This research has been funded by Universidad Panamericana through the grant &#8220;Fomento a la Investigaci&#243;n UP 2018&#8221;, under project code UP-CI-2018-ING-MX-04.

All methods described here have been approved by the Research Committee of the School of Engineering of Universidad Panamericana.
NOTE: This methodology is based on the configuration of the specific types of sensors, machine-learning methods and procedures in order to configure a simple, fast and multimodal fall detection and human activity recognition system. Due to this, the following protocol is divided in phases: (1) database creation (2) data analysis (3) system simplification and (4) evaluation.
1. Database creation
<ol>
	<li>Set up the data acquisition system. This will collect all data from subjects and store the information in a retrieval database.
	<ol>
		<li>Select the types of wearable sensors, ambient sensors and vision-based devices required as sources of information. Assign an ID for each source of information, the number of channels per source, the technical specifications and the sampling rate of each of them.</li>
		<li>Connect all sources of information (i.e., wearables and ambient sensors, and vision-based devices) to a central computer or a distributed computer system:
		<ol>
			<li>Verify that wired-based devices are connected properly to one client computer. Verify that wireless-based devices are fully charged. Consider that low battery might impact wireless connections or sensor values. Moreover, intermittent or lost connections will increase loss of data.</li>
		</ol>
		</li>
		<li>Set up each of the devices to retrieve data.</li>
		<li>Set up the data acquisition system for storing data on the cloud. Due to the large amount of data to be stored, cloud computing is considered in this protocol.</li>
		<li>Validate that the data acquisition system fulfills data synchronization and data consistency20 properties. This maintains the integrity of data storage from all the sources of information. It might require new approaches in data synchronization. For example, see Pe&#241;afort-Asturiano et al.20.
		<ol>
			<li>Start collecting some data with the sources of information and store data in a preferred system. Include timestamps in all data.</li>
			<li>Query the database and determine if all sources of information are collected at the same sample rates. If done properly, go to Step 1.1.6. Otherwise, perform up-sampling or down-sampling using criteria reported in Pe&#241;afort-Asturiano, et al.20.</li>
		</ol>
		</li>
		<li>Set up the environment (or laboratory) by considering the conditions required and the restrictions imposed by the goal of the system. Set conditions for impact force attenuation in the simulated falls as compliant flooring systems suggested in Lachance, et al.23 to ensure participants safety.
		<ol>
			<li>Use a mattress or any other compliant flooring system and place it at the center of the environment (or laboratory).</li>
			<li>Keep all objects away from the mattress to give at least one meter of safe space all around. If required, prepare personal protective equipment for participants (e.g., gloves, cap, goggles, knee support, etc.). 
			NOTE: The protocol can be paused here.</li>
		</ol>
		</li>
	</ol>
	</li>
	<li>Determine the human activities and falls that the system will detect after configuration. It is important to have in mind the purpose of the fall detection and human activity recognition system, as well as the target population.
	<ol>
		<li>Define the goal of the fall detection and human activity recognition system. Write it down in a planning sheet. For this case study, the goal is to classify the types of human falls and activities performed in an indoor daily basis of elderly people.</li>
		<li>Define the target population of the experiment in accordance with the goal of the system. Write it down in the planning sheet. In the study, consider elderly people as the target population.</li>
		<li>Determine the type of daily activities. Include some non-fall activities that look like falls in order to improve real fall detection. Assign an ID for all of them and describe them as detailed as possible. Set the time period for each activity to be executed. Write all this information down in the planning sheet.</li>
		<li>Determine the type of human falls. Assign an ID for all of them and describe them as detailed as possible. Set the time period for each fall to be executed . Consider if the falls will be self-generated by the subjects or generated by others (e.g., pushing the subject). Write all this information down in the planning sheet.</li>
		<li>In the planning sheet, write down the sequences of activities and falls that a subject will perform. Specify the period of time, the number of trials per activity/fall, the description to perform the activity/fall, and the activity/fall IDs. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
	<li>Select the relevant subjects to the study that will execute the sequences of activities and falls. Falls are rare events to catch in real life and usually occur to old persons. Nevertheless, for safety reasons, do not include elderly and impaired people in fall simulation under medical advice. Stunts have been used to avoid injuries22.
	<ol>
		<li>Determine the gender, age range, weight and height of the subjects. Define any impairment conditions required. Also, define the minimum number of subjects required for the experiment.</li>
		<li>Randomly select the set of subjects required, following the conditions stated in the previous step. Use a call for volunteers to recruit them. Fulfill all ethical guidelines applicable from the institution and country, as well as any international regulation when experimenting with humans. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
	<li>Retrieve and store data from subjects. This information will be useful for further experimental analysis. Complete the following steps under supervision by a clinical expert or a responsible researcher.
	<ol>
		<li>Start collecting data with the data acquisition system configured in Step 1.1.</li>
		<li>Ask each of the subjects to perform the sequences of activities and falls declared in Step 1.2. Clearly save the timestamps of the start and end of each activity/fall. Verify that data from all sources of information is saved on the cloud.</li>
		<li>If the activities were not properly done or there were issues with devices (e.g., lost connection, low battery, intermittent connection), discard the samples and repeat Step 1.4.1 until no device issues are found. Repeat Step 1.4.2 for each trial, per subject, declared in the sequence of Step 1.2. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
	<li>Pre-process all data acquired. Apply up-sampling and down-sampling for each of the sources of information. See details about pre-processing data for fall detection and human activity recognition in Mart&#237;nez-Villase&#241;or et al.21. 
	NOTE: The protocol can be paused here.</li>
</ol>
2. Data Analysis
<ol>
	<li>Select the mode of data treatment. Select Raw Data if the data stored in the database will be used outright (i.e., using deep learning for automatic feature extraction) and go to Step 2.2. Select Feature Data if feature extraction will be used for further analysis and go to Step 2.3.</li>
	<li>For Raw Data, no extra steps are required so go to Step 2.5.</li>
	<li>For Feature Data, extract features from the raw data.
	<ol>
		<li>Segment raw data in time windows. Determine and fix the time window length (e.g., frames of one-second size). In addition, determine if these time windows will be overlapping or not. A good practice is to choose 50% overlapping.</li>
		<li>Extract features from each segment of data. Determine the set of temporal and frequential features to be extracted from the segments. See Mart&#237;nez-Villase&#241;or et al.21 for common feature extraction.</li>
		<li>Save the feature extraction data set on the cloud, in an independent database.</li>
		<li>If different time windows will be selected, repeat Steps 2.3.1 to 2.3.3, and save each feature data set in independent databases. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
	<li>Select the most important features extracted and reduce the feature data set. Apply some commonly used feature selection methods (e.g., univariate selection, principal components analysis, recursive feature elimination, feature importance, correlation matrix, etc.).
	<ol>
		<li>Select a feature selection method. Here, we used feature importance.</li>
		<li>Use each feature to train a given model (we employed RF) and measure the accuracy (see Equation 1).</li>
		<li>Rank the features by sorting in order of the accuracy.</li>
		<li>Select the most important features. Here, we used the best ranked first ten features. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
	<li>Select a machine learning classification method and train a model. There are well-known machine learning methods16,17,18,21, such as: support vector machines (SVM), random forest (RF), multilayer perceptron (MLP) and k-nearest neighbors (KNN), among many others.
	<ol>
		<li>Optionally, if a deep learning approach is selected, then consider21: convolutional neural networks (CNN), long short-term memory neural networks (LSTM), among others.</li>
		<li>Select a set of machine learning methods. Here, we used the following methods: SVM, RF, MLP and KNN.</li>
		<li>Fix the parameters of each of the machine learning methods, as suggested in literature21.</li>
		<li>Create a combined feature data set (or raw data set) using the independent feature data sets (or raw data sets), to combine types of sources of information. For example, if a combination of one wearable sensor and one camera is required, then combine the feature data sets from each of these sources.</li>
		<li>Split the feature data set (or raw data set) in training and testing sets. A good choice is to randomly divide 70% for training and 30% for testing.</li>
		<li>Run a k-fold cross-validation21 using the feature data set (or raw data set), for each machine learning method. Use a common metric of evaluation, like accuracy (see Equation 1) to select the best model trained per method. Leave-one subject-out (LOSO) experiments3 are also recommend.
		<ol>
			<li>Open the training feature data set (or raw data set) in the preferred programming language software. Python is recommended. For this step, use pandas library to read a CSV file as follows: 
			training_set = pandas.csv(&#60;filename.csv&#62;).</li>
			<li>Split the feature data set (or raw data set) in pairs of inputs-outputs. For example, use Python to declare the x-values (inputs) and the y-values (outputs): 
			training_set_X = training_set.drop(&#8216;tag&#8217;,axis=1), training_set_Y = training_set.tag 
			where tag represents the column of the feature data set that includes the target values.</li>
			<li>Select one machine learning method and set the parameters. For example, use SVM in Python with the library sklearn like the following command: 
			classifier = sklearn.SVC(kernel = &#8216;poly&#8217;) 
			in which the kernel function is selected as polynomial.</li>
			<li>Train the machine learning model. For example, use the above classifier in Python to train the SVM model: 
			classifier.fit(training_set_X,training_set_Y).</li>
			<li>Compute the estimates values of the model using the testing feature data set (or the raw data set). For example, use the estimate function in Python as follows: estimates = classifier.predict(testing_set_X) where testing_set_X represents the x-values of the testing set.</li>
			<li>Repeat Steps 2.5.6.1 to 2.5.6.5, the number of times k specified in the k-fold cross validation (or the number of times required for the LOSO approach).</li>
			<li>Repeat Steps 2.5.6.1 to 2.5.6.6 for each machine learning model selected. 
			NOTE: The protocol can be paused here.</li>
		</ol>
		</li>
		<li>Compare the machine learning methods by testing the selected models with the testing data set. Other metrics of evaluation can be used: accuracy (Equation 1), precision (Equation 2), sensitivity (Equation 3), specificity (Equation 4) or F1-score (Equation 5), where TP are the true positives, TN are the true negatives, FP are the false positives and FN are the false negatives. 
		<img alt="Equation 1" src="/files/ftp_upload/60361/60361equ01.jpg" /> 
		<img alt="Equation 2" src="/files/ftp_upload/60361/60361equ02.jpg" /> 
		<img alt="Equation 3" src="/files/ftp_upload/60361/60361equ03.jpg" /> 
		<img alt="Equation 4" src="/files/ftp_upload/60361/60361equ04.jpg" /> 
		<img alt="Equation 5" src="/files/ftp_upload/60361/60361equ05.jpg" /></li>
		<li>Use other beneficial performance metrics such as the confusion matrix9 to evaluate the classification task of the machine learning models, or a decision-independent precision-recall9 (PR) or receiver operating characteristic9 (ROC) curves. In this methodology, recall and sensitivity are considered equivalent.</li>
		<li>Use qualitative features of the machine learning models to compare among them, such as: ease of machine learning interpretation; real-time performance; limited resources of time, memory and processing computing; and ease of machine learning deployment in edge devices or embedded systems.</li>
		<li>Select the best machine learning model using the information from: The quality metrics (Equations 1&#8211;5), the performance metrics and the qualitative features of machine learning feasibility of Steps 2.5.6, 2.5.7 and 2.5.8. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
</ol>
3. System simplification
<ol>
	<li>Select the suitable placements of sources of information. Sometimes, it is necessary to determine the best placement of sources of information (e.g., which location of a wearable sensor is better).
	<ol>
		<li>Determine the subset of sources of information that will be analyzed. For example, if there are five wearable sensors in the body and just one has to be selected as the best sensor placed, each of these sensors will be part of the subset.</li>
		<li>For each source of information in this subset, create a separate data set and store it separately. Keep in mind that this data set could be either the previous feature data set or the raw data set. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
	<li>Select a machine learning classification method and train a model for one source of information placement. Complete Steps from 2.5.1 to 2.5.6 using each of the data sets created in Step 3.1.2. Detect the most suitable source of information placement by ranking. For this case study, we use the following methods: SVM, RF, MLP and KNN. 
	Note: The protocol can be paused here.</li>
	<li>Select the suitable placements in a multimodal approach if a combination of two or more sources of information are required for the system (e.g., combination of one wearable sensor and one camera). In this case study, use waist-wearable sensor and camera 1 (lateral view) as the modalities.
	<ol>
		<li>Select the best source of information of each modality in the system and create a combined feature data set (or raw data set) using the independent data sets of these sources of information.</li>
		<li>Select a machine learning classification method and train a model for these combined sources of information. Complete Steps 2.5.1 to 2.5.6 using the combined feature data set (or raw data set). In this study, use the following methods: SVM, RF, MLP and KNN. 
		NOTE: The protocol can be paused here.</li>
	</ol>
	</li>
</ol>
4. Evaluation
<ol>
	<li>Prepare a new data set with users in more realistic conditions. Use only the sources of information selected in the previous step. Preferable, implement the system in the target group (e.g., elderly people). Collect data in longer periods of time.
	<ol>
		<li>Optionally if the target group is used only, create a selection group protocol including the terms of exclusion (e.g., any physical or psychological impairment) and stop criteria prevention (e.g., detect any physical injury during the trials; suffering nausea, dizziness and/or vomiting; fainting). Consider also ethical concerns and data privacy issues.</li>
	</ol>
	</li>
	<li>Evaluate the performance of the fall detection and human activity recognition system developed so far. Use Equations 1&#8211;5 to determine the accuracy and predictive power of the system, or any other performance metrics.</li>
	<li>Discuss about the findings on the experimental results.</li>
</ol>

<ol><li>United Nations. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((World+Population+Prospects%3A+The+2017+Revision%2C+Key+Findings+and+Advance+Tables%5BTitle%5D)+AND+%22United+Nations.+Department+of+Economic+and+Social+Affairs%2C+Population+Division%22%5BJournal%5D)">World Population Prospects: The 2017 Revision, Key Findings and Advance Tables.</a> United Nations. Department of Economic and Social Affairs, Population Division. , ESA/P/WP/248 (2017).</li>
<li>World Health Organization. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Ageing%2C+and+Life+Course+Unit%5BTitle%5D)+AND+%22WHO+Global+Report+on+Falls+Prevention+in+Older+Age%22%5BJournal%5D)">Ageing, and Life Course Unit.</a> WHO Global Report on Falls Prevention in Older Age. , (2008).</li>
<li>Igual, R., Medrano, C., Plaza, I. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Challenges%2C+Issues+and+Trends+in+Fall+Detection+Systems%5BTitle%5D)+AND+%22Biomedical+Engineering+Online%22%5BJournal%5D)">Challenges, Issues and Trends in Fall Detection Systems.</a> Biomedical Engineering Online. 12 (1), 66(2013).</li>
<li>Noury, N., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Fall+Detection-Principles+and+Methods%5BTitle%5D)+AND+%222007+29th+Annual+International+Conference+of+the+IEEE+Engineering+in+Medicine+and+Biology+Society%22%5BJournal%5D)">Fall Detection-Principles and Methods.</a> 2007 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. , 1663-1666 (2007).</li>
<li>Mubashir, M., Shao, L., Seed, L. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((A+Survey+on+Fall+Detection%3A+Principles+and+Approaches%5BTitle%5D)+AND+%22Neurocomputing%22%5BJournal%5D)">A Survey on Fall Detection: Principles and Approaches.</a> Neurocomputing. 100, 144-152 (2002).</li>
<li>Perry, J. T., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Survey+and+Evaluation+of+Real-Time+Fall+Detection+Approaches%5BTitle%5D)+AND+%22Proceedings+of+the+6th+International+Symposium+High-Capacity+Optical+Networks+and+Enabling+Technologies%22%5BJournal%5D)">Survey and Evaluation of Real-Time Fall Detection Approaches.</a> Proceedings of the 6th International Symposium High-Capacity Optical Networks and Enabling Technologies. , 158-164 (2009).</li>
<li>Xu, T., Zhou, Y., Zhu, J. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((New+Advances+and+Challenges+of+Fall+Detection+Systems%3A+A+Survey%5BTitle%5D)+AND+%22Applied+Sciences%22%5BJournal%5D)">New Advances and Challenges of Fall Detection Systems: A Survey.</a> Applied Sciences. 8 (3), 418(2018).</li>
<li>Rougier, C., Meunier, J., St-Arnaud, A., Rousseau, J. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Robust+Video+Surveillance+for+Fall+Detection+Based+on+Human+Shape+Deformation%5BTitle%5D)+AND+%22IEEE+Transactions+on+Circuit+Systems+for+Video+Technologies%22%5BJournal%5D)">Robust Video Surveillance for Fall Detection Based on Human Shape Deformation.</a> IEEE Transactions on Circuit Systems for Video Technologies. 21, 611-622 (2011).</li>
<li>Bulling, A., Blanke, U., Schiele, B. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((A+Tutorial+on+Human+Activity+Recognition+Using+Body-Worn+Inertial+Sensors%5BTitle%5D)+AND+%22ACM+Computing+Surveys%22%5BJournal%5D)">A Tutorial on Human Activity Recognition Using Body-Worn Inertial Sensors.</a> ACM Computing Surveys. 46 (3), 33(2014).</li>
<li>Kwolek, B., Kepski, M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Human+Fall+Detection+on+Embedded+Platform+Using+Depth+Maps+and+Wireless+Accelerometer%5BTitle%5D)+AND+%22Computational+Methods+and+Programs+in+Biomedicine%22%5BJournal%5D)">Human Fall Detection on Embedded Platform Using Depth Maps and Wireless Accelerometer.</a> Computational Methods and Programs in Biomedicine. 117, 489-501 (2014).</li>
<li>Ofli, F., Chaudhry, R., Kurillo, G., Vidal, R., Bajcsy, R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Berkeley+MHAD%3A+A+Comprehensive+Multimodal+Human+Action+Database%5BTitle%5D)+AND+%22Proceedings+of+the+2013+IEEE+Workshop+on+Applications+of+Computer+Vision%22%5BJournal%5D)">Berkeley MHAD: A Comprehensive Multimodal Human Action Database.</a> Proceedings of the 2013 IEEE Workshop on Applications of Computer Vision. , 53-60 (2013).</li>
<li>Dovgan, E., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Intelligent+Elderly-Care+Prototype+for+Fall+and+Disease+Detection%5BTitle%5D)+AND+%22Slovenian+Medical+Journal%22%5BJournal%5D)">Intelligent Elderly-Care Prototype for Fall and Disease Detection.</a> Slovenian Medical Journal. 80, 824-831 (2011).</li>
<li>Santoyo-Ramón, J., Casilari, E., Cano-García, J. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Analysis+of+a+Smartphone-Based+Architecture+With+Multiple+Mobility+Sensors+for+Fall+Detection+With+Supervised+Learning%5BTitle%5D)+AND+%22Sensors%22%5BJournal%5D)">Analysis of a Smartphone-Based Architecture With Multiple Mobility Sensors for Fall Detection With Supervised Learning.</a> Sensors. 18 (4), 1155(2018).</li>
<li>Özdemir, A. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((An+Analysis+on+Sensor+Locations+of+the+Human+Body+for+Wearable+Fall+Detection+Devices%3A+Principles+and+Practice%5BTitle%5D)+AND+%22Sensors%22%5BJournal%5D)">An Analysis on Sensor Locations of the Human Body for Wearable Fall Detection Devices: Principles and Practice.</a> Sensors. 16 (8), 1161(2016).</li>
<li>Ntanasis, P., Pippa, E., Özdemir, A. T., Barshan, B., Megalooikonomou, V. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Investigation+of+Sensor+Placement+for+Accurate+Fall+Detection%5BTitle%5D)+AND+%22International+Conference+on+Wireless+Mobile+Communication+and+Healthcare%22%5BJournal%5D)">Investigation of Sensor Placement for Accurate Fall Detection.</a> International Conference on Wireless Mobile Communication and Healthcare. , 225-232 (2016).</li>
<li>Bagala, F., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Evaluation+of+Accelerometer-Based+Fall+Detection+Algorithms+on+Real-World+Falls%5BTitle%5D)+AND+%22PLoS+One%22%5BJournal%5D)">Evaluation of Accelerometer-Based Fall Detection Algorithms on Real-World Falls.</a> PLoS One. 7, 37062(2012).</li>
<li>Bourke, A. K., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Assessment+of+Waist-Worn+Tri-Axial+Accelerometer+Based+Fall-detection+Algorithms+Using+Continuous+Unsupervised+Activities%5BTitle%5D)+AND+%22Proceedings+of+the+Annual+International+Conference+of+the+IEEE+Engineering+in+Medicine+and+Biology+Society%22%5BJournal%5D)">Assessment of Waist-Worn Tri-Axial Accelerometer Based Fall-detection Algorithms Using Continuous Unsupervised Activities.</a> Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society. , 2782-2785 (2010).</li>
<li>Kerdegari, H., Samsudin, K., Ramli, A. R., Mokaram, S. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Evaluation+of+Fall+Detection+Classification+Approaches%5BTitle%5D)+AND+%224th+International+Conference+on+Intelligent+and+Advanced+Systems%22%5BJournal%5D)">Evaluation of Fall Detection Classification Approaches.</a> 4th International Conference on Intelligent and Advanced Systems. , 131-136 (2012).</li>
<li>Alazrai, R., Mowafi, Y., Hamad, E. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((A+Fall+Prediction+Methodology+for+Elderly+Based+on+a+Depth+Camera%5BTitle%5D)+AND+%2237th+Annual+International+Conference+of+the+IEEE+Engineering+in+Medicine+and+Biology+Society%22%5BJournal%5D)">A Fall Prediction Methodology for Elderly Based on a Depth Camera.</a> 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. , 4990-4993 (2015).</li>
<li>Peñafort-Asturiano, C. J., Santiago, N., Núñez-Martínez, J. P., Ponce, H., Martínez-Villaseñor, L. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Challenges+in+Data+Acquisition+Systems%3A+Lessons+Learned+from+Fall+Detection+to+Nanosensors%5BTitle%5D)+AND+%222018+Nanotechnology+for+Instrumentation+and+Measurement%22%5BJournal%5D)">Challenges in Data Acquisition Systems: Lessons Learned from Fall Detection to Nanosensors.</a> 2018 Nanotechnology for Instrumentation and Measurement. , 1-8 (2018).</li>
<li>Martínez-Villaseñor, L., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((UP-Fall+Detection+Dataset%3A+A+Multimodal+Approach%5BTitle%5D)+AND+%22Sensors%22%5BJournal%5D)">UP-Fall Detection Dataset: A Multimodal Approach.</a> Sensors. 19 (9), 1988(2019).</li>
<li>Rantz, M., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Falls%2C+Technology%2C+and+Stunt+Actors%3A+New+approaches+to+Fall+Detection+and+Fall+Risk+Assessment%5BTitle%5D)+AND+%22Journal+of+Nursing+Care+Quality%22%5BJournal%5D)">Falls, Technology, and Stunt Actors: New approaches to Fall Detection and Fall Risk Assessment.</a> Journal of Nursing Care Quality. 23 (3), 195-201 (2008).</li>
<li>Lachance, C., Jurkowski, M., Dymarz, A., Mackey, D. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Compliant+Flooring+to+Prevent+Fall-Related+Injuries%3A+A+Scoping+Review+Protocol%5BTitle%5D)+AND+%22BMJ+Open%22%5BJournal%5D)">Compliant Flooring to Prevent Fall-Related Injuries: A Scoping Review Protocol.</a> BMJ Open. 6 (8), 011757(2016).</li>
</ol>

The authors have nothing to disclose.

It is common to encounter challenges due to synchronization, organization and data inconsistency problems20 when a dataset is created.
Synchronization 
In the acquisition of data, synchronization problems arise given that multiple sensors commonly work at different sampling rates. Sensors with higher frequencies collect more data than those with lower frequencies. Thus, data from different sources will not be paired correctly. Even if sensors run at the same sampling rates, it is possible that data will not be aligned. In this regard, the following recommendations might help to handle these synchronization problems20: (i) register timestamp, subject, activity and trial in each data sample obtained from the sensors; (ii) the most consistent and less frequent source of information has to be used as reference signal for synchronization; and (iii) use automatic or semi-automatic procedures to synchronize video recordings that manual inspection would be impractical.
Data pre-processing 
Data pre-processing must also be done, and critical decisions influence this process: (a) determine the methods for data storage and data representation of multiple and heterogeneous sources (b) decide the ways to store data in the local host or on the cloud (c) select the organization of data, including the file names and folders (d) handle missing values of data as well as redundancies found in the sensors, among others. In addition, for the data cloud, local buffering is recommended when possible to mitigate loss of data at the uploading time.
Data inconsistency 
Data inconsistency is common between trials finding variations in data sample sizes. These issues are related to data acquisition in wearable sensors. Brief interruptions of data acquisition and data collision from multiple sensors leads to data inconsistencies. In these cases, inconsistency detection algorithms are important to handle online failure in sensors. It is important to highlight that wireless-based devices should be monitored frequently throughout the experiment. Low battery might impact connectivity and result in loss of data.
Ethical 
Consent to participate and ethical approval are mandatory in every type of experimentation where people are involved.
Regarding the limitations of this methodology, it is important to notice that it is designed for approaches that consider different modalities for data collection. The systems can include wearable, ambient and/or vision sensors. It is suggested to consider the power consumption of devices and the lifetime of batteries in wireless-based sensors, due to the issues such as loss of data collection, diminishing connectivity and power consumption in the whole system. Moreover, this methodology is intended for systems that use machine learning methods. An analysis of the selection of these machine learning models should be done beforehand. Some of these models could be accurate, but highly time and energy consuming. A trade-off between accurate estimation and limited resource availability for computing in machine learning models must be taken into consideration. It is also important to observe that, in data collection of the system, the activities were conducted in the same order; also, trials were performed in the same sequence. For safety reasons, a protective mattress was used for subjects to fall onto. In addition, the falls were self-initiated. This is an important difference between simulated and real falls, which generally occur towards hard materials. In that sense, this dataset recorded falls with an intuitive reaction trying not to fall. Moreover, there are some differences between real falls in elderly or impaired people and the simulation falls; and these must be taken into account when designing a new fall detection system. This study was focused on young people without any impairments, but it is remarkable to say that the selection of subjects should be aligned to the goal of the system and the target population who will use it.
From the related works described above10,11,12,13,14,15,16,17,18, we can observe that there are authors that use multimodal approaches focusing in obtaining robust fall detectors or focus on placement or performance of the classifier. Hence, they only address one or two of the design issues for fall detection. Our methodology allows solving simultaneously three of the main design problems of a fall detection system.
For future work, we suggest designing and implementing a simple multimodal fall detection system based on the findings obtained following this methodology. For real-world adoption, transfer learning, hierarchical classification and deep learning approaches should be used for developing more robust systems. Our implementation did not consider qualitative metrics of the machine learning models, but real-time and limited computing resources have to be taken into account for further development of human fall and activity detection/recognition systems. Lastly, in order to improve our dataset, tripping or almost falling activities and real-time monitoring of volunteers during their daily life can be considered.

Since the world phenomenon of population aging1, fall prevalence has increased and is actually considered a major health problem2. When a fall occurs, people require immediate attention in order to reduce negative consequences. Fall detection systems can reduce the amount of time in which a person receives medical attention sending an alert when a fall occurs.
There are various categorizations of fall detection systems3. Early works4 classify fall detection systems by their method of detection, roughly analytical methods and machine learning methods. More recently, other authors3,5,6 have considered data acquisition sensors as the main feature to classify fall detectors. Igual et al.3 divides fall detection systems into context-aware systems, that include vision and ambient-sensor based approaches, and wearable device systems. Mubashir et al.5 classifies fall detectors into three groups based on the devices used for data acquisition: wearable devices, ambience sensors, and vision-based devices. Perry et al.6 considers methods for measuring acceleration, methods for measuring acceleration combined with other methods, and methods not measuring acceleration. From these surveys, we can determine that sensors and methods are the main elements to classify the general research strategy.
Each of the sensors has weaknesses and strengths discussed in Xu et al.7. Vision-based approaches mainly use normal cameras, depth sensor cameras, and/or motion capture systems. Normal web cameras are low cost and easy to use, but they are sensitive to environmental conditions (light variation, occlusion, etc.), can only be used in a reduced space, and have privacy issues. Depth cameras, such as the Kinect, provide full-body 3D motion7 and are less affected by lighting conditions than normal cameras. However, approaches based on the Kinect are not as robust and reliable. Motion capture systems are more expensive and difficult to use.
Approaches based on accelerometer devices and smart phones/watches with built-in accelerometers are very commonly used for fall detection. The main drawback of these devices is that they have to be worn for long periods. Discomfort, obtrusiveness, body placement and orientation are design issues to be solved in these approaches. Although smartphones and smart watches are less obtrusive devices that sensors, older people often forget or do not always wear these devices. Nevertheless, the advantage of these sensors and devices is that they can be used in many rooms and/or outdoors.
Some systems use sensors placed around the environment to recognize falls/activities, so people do not have to wear the sensors. However, these sensors are also limited to the places where they are deployed8 and are sometimes difficult to install. Recently, multimodal fall detection systems include different combinations of vision, wearable and ambient sensors in order to gain more precision and robustness. They can also overcome some of the single sensor limitations.
The methodology used for fall detection is closely related with human activity recognition chain (ARC) presented by Bulling et al.9, which consists of stages for data acquisition, signal preprocessing and segmentation, feature extraction and selection, training and classification. Design issues must be solved for each of these stages. Different methods are used in each stage.
We present a methodology based on multimodal sensors to configure a simple, comfortable and fast human fall and human activity detection/recognition system. The goal is to build a system for accurate fall detection that can be easily implemented and adopted. The proposed novel methodology is based on ARC, but it adds some important phases to perform a deep analysis of the following issues in order to simplify the system: (a) select which sensors or combination of sensors are to be used in a simple fall detection system; (b) determine the best placement of the information sources ; and (c) select the most suitable machine learning classification method for fall detection and human activity recognition to create a simple system.
There are some related works in literature that address one or two of the above-mentioned design issues, but to our knowledge, there is no work that focuses on a methodology to overcome all of these problems.
Related works use multimodal approaches for fall detection and human activity recognition10,11,12 in order to gain robustness and increase precision. Kwolek et al.10 proposed the design and implementation of a fall detection system based on accelerometric data and depth maps. They designed an interesting methodology in which a three-axis accelerometer is implemented to detect a potential fall as well as the person&#8217;s motion. If the acceleration measure exceeds a threshold, the algorithm extracts a person differencing the depth map from the online updated depth reference map. An analysis of depth and accelerometer combinations was made using a support vector machine classifier.
Ofli et al.11 presented a Multimodal Human Action Database (MHAD) in order to provide a testbed for new human activity recognition systems. The dataset is important since the actions were gathered simultaneously using 1 optical motion capture system, 4 multi-view cameras, 1 Kinect system, 4 microphones, and 6 wireless accelerometers. The authors presented results for each modality: the Kinect, the mocap, the accelerometer, and the audio.
Dovgan et al.12 proposed a prototype for detecting anomalous behavior, including falls, in the elderly. They designed tests for three sensor systems in order to find the most appropriate equipment for fall and unusual-behavior detection. The first experiment consists of data from a smart sensor system with 12 tags attached to the hips, knees, ankles, wrists, elbows and shoulders. They also created a test dataset using one Ubisense sensor system with four tags attached to the waist, chest and both ankles, and one Xsens accelerometer. In a third experiment, four subjects only use the Ubisense system while performing 4 types of falls, 4 health problems as anomalous behavior and different activity of daily living (ADL).
Other works in literature13,14,15 address the problem of finding the best placement of sensors or devices for fall detection comparing the performance of various combinations of sensors with several classifiers. Santoyo et al.13 presented a systematic assessment evaluating the importance of the location of 5 sensors for fall detection. They compared the performance of these sensor combinations using k-nearest neighbors (KNN), support vector machines (SVM), na&#239;ve Bayes (NB) and decision tree (DT) classifiers. They conclude that the location of the sensor on the subject has an important influence on the fall detector performance independent of the classifier used.
A comparison of wearable sensor placements on the body for fall detection was presented by &#214;zdemir14. In order to determine sensor placement, the author analyzed 31 sensor combinations of the following positions: head, waist, chest, right wrist, right ankle and right thigh. Fourteen volunteers performed 20 simulated falls and 16 ADL. He found that the best performance was obtained when a single sensor is positioned on the waist from these exhaustive combination experiments. Another comparison was presented by Ntanasis15 using &#214;zdemir&#8217;s dataset. The authors compared single positions on the head, chest, waist, wrist, ankle and thigh using the following classifiers: J48, KNN, RF, random committee (RC) and SVM.
Benchmarks of the performance of different computational methods for fall detection can also be found in literature16,17,18. Bagala et al.16 presented a systematic comparison to benchmark the performance of thirteen fall detection methods tested on real falls. They only considered algorithms based on accelerometer measurements placed on the waist or trunk. Bourke et al.17 evaluated the performance of five analytical algorithms for fall detection using a dataset of ADLs and falls based on accelerometer readings. Kerdegari18 made also a comparison of the performance of different classification models for a set of recorded acceleration data. The algorithms used for fall detection were zeroR, oneR, NB, DT, multilayer perceptron and SVM.
A methodology for fall detection was proposed by Alazrai et al.18 using motion pose geometric descriptor to construct an accumulated histogram-based representation of human activity. They evaluated the framework using a dataset collected with Kinect sensors.
In summary, we found multimodal fall detection related works10,11,12 that compare the performance of different combinations of modalities. Some authors address the problem of finding the best placement of sensors13,14,15, or combinations of sensors13 with several classifiers13,15,16 with multiple sensors of the same modality and accelerometers. No work was found in literature that address placement, multimodal combinations and classifier benchmark at the same time.

<table><tbody><tr><td>Inertial measurement wearable sensor</td><td>Mbientlab</td><td>MTH-MetaTracker</td><td>Tri-axial accelerometer, tri-axial gyroscope and light intensity wearable sensor.</td></tr><tr><td>Electroencephalograph brain sensor helmet MindWave</td><td>NeuroSky</td><td>80027-007</td><td>Raw brainwave signal with one forehand sensor.</td></tr><tr><td>LifeCam Cinema video camera</td><td>Microsoft</td><td>H5D-00002</td><td>2D RGB camera with USB cable interface.</td></tr><tr><td>Infrared sensor</td><td>Alean</td><td>ABT-60</td><td>Proximity sensor with normally closed relay.</td></tr><tr><td>Bluetooth dongle</td><td>Mbientlab</td><td>BLE</td><td>Dongle for Bluetooth connection between the wearable sensors and a computer.</td></tr><tr><td>Raspberry Pi</td><td>Raspberry</td><td>Version 3 Model B</td><td>Microcontroller for infrared sensor acquisition and computer interface.</td></tr><tr><td>Personal computer</td><td>Dell</td><td></td><td>Intel Xeon E5-2630 v4 @2.20 GHz, RAM 32GB</td></tr></tbody></table>

design and analysis for fall detection system simplification

This paper presents a methodology based on multimodal sensors to configure a simple, comfortable and fast fall detection and human activity recognition system that can be easily implemented and adopted. The methodology is based on the configuration of specific types of sensors, machine-learning methods and procedures. The protocol is divided into four phases: (1) database creation (2) data analysis (3) system simplification and (4) evaluation. Using this methodology, we created a multimodal database for fall detection and human activity recognition, namely UP-Fall Detection. It comprises data samples from 17 subjects that perform 5 types of falls and 6 different simple activities, during 3 trials. All information was gathered using 5 wearable sensors (tri-axis accelerometer, gyroscope and light intensity), 1 electroencephalograph helmet, 6 infrared sensors as ambient sensors, and 2 cameras in lateral and front viewpoints. The proposed novel methodology adds some important stages to perform a deep analysis of the following design issues in order to simplify a fall detection system: a) select which sensors or combination of sensors are to be used in a simple fall detection system, b) determine the best placement of the sources of information, and c) select the most suitable machine learning classification method for fall and human activity detection and recognition. Even though some multimodal approaches reported in literature only focus on one or two of the above-mentioned issues, our methodology allows simultaneously solving these three design problems related to a human fall and activity detection and recognition system.

Watch this Scientific Journal Video about Design and Analysis for Fall Detection System Simplification at JoVE.com

Design and Analysis for Fall Detection System Simplification

We present a methodology based on multimodal sensors to configure a simple, comfortable and fast fall detection and human activity recognition system. The goal is to build a system for accurate fall detection that can be easily implemented and adopted.

This paper presents a methodology based on multimodal sensors to configure a simple, comfortable and fast fall detection and human activity recognition ...

design-and-analysis-for-fall-detection-system-simplification

Research

JoVE Journal

Bioengineering

10.6K Views.  Universidad Panamericana. Onze methodologie voegt belangrijke fasen, selectie van sensorcombinatie, plaatsing en classificatie, toe om het valdetectiesysteem te vereenvoudigen met een diepe analyse. Er zijn eerdere werken die een aantal problemen met het ontwerp van valdetectie aanpakken, maar er is geen werk dat zich richt op holistische methodologie voor het overwinnen van al deze problemen. Deze methodologie kan ook worden gebruikt voor menselijke activiteit erkenning in begeleid wonen, sport prestatie-evaluatie, f...

Video: Design and Analysis for Fall Detection System Simplification

A Precise and Autonomous System for the Detection of Insect Emergence Patterns

Existing systems to measure insect emergence patterns have limitations; they are only partially automated and are limited in the maximum number of emerging insects they can detect. In order to obtain precise measurement of insect emergence, it is necessary for systems to be semi-automated and able to measure large numbers of emerging insects. We addressed these issues by designing and building a system that is automated and can measure emergence of up to 1200 insects. We modified the existing &#34;falling-ball&#34; system using Arduino microcontrollers to automate data collection and expand the sample size through multiple data channels. Multiple data channels enable the user to not only increase their sample size, but also allows for multiple treatments to be run simultaneously in a single experiment. Furthermore, we created an R script to automatically visualize the data as a bubble plot, while also calculating the median day and time of emergence. The current system was designed using 3D printing so the user can modify the system to be adjusted for different species of insects. The goal of this protocol is to investigate important questions in chronobiology and stress physiology, using this precise and automated system to measure insect emergence patterns.

Measurement of insect emergence patterns requires precision. Existing systems are only semi-automated and sample size is limited. We addressed these issues by designing a system using microcontrollers to precisely measure the time of emergence of large numbers of emerging insects.

Existing systems to measure insect emergence patterns have limitations; they are only partially automated and are limited in the maximum number of ...

Environment

The Modular Design and Production of an Intelligent Robot Based on a Closed-Loop Control Strategy

Intelligent robots are part of a new generation of robots that are able to sense the surrounding environment, plan their own actions and eventually reach their targets. In recent years, reliance upon robots in both daily life and industry has increased. The protocol proposed in this paper describes the design and production of a handling robot with an intelligent search algorithm and an autonomous identification function.
First, the various working modules are mechanically assembled to complete the construction of the work platform and the installation of the robotic manipulator. Then, we design a closed-loop control system and a four-quadrant motor control strategy, with the aid of debugging software, as well as set steering gear identity (ID), baud rate and other working parameters to ensure that the robot achieves the desired dynamic performance and low energy consumption. Next, we debug the sensor to achieve multi-sensor fusion to accurately acquire environmental information. Finally, we implement the relevant algorithm, which can recognize the success of the robot&#39;s function for a given application.
The advantage of this approach is its reliability and flexibility, as the users can develop a variety of hardware construction programs and utilize the comprehensive debugger to implement an intelligent control strategy. This allows users to set personalized requirements based on their needs with high efficiency and robustness.

We present a protocol on modular design and production of intelligent robots to help scientific and technical workers design intelligent robots with special production tasks based on personal needs and individualized design.

Intelligent robots are part of a new generation of robots that are able to sense the surrounding environment, plan their own actions and eventually reach ...

Engineering

Using a Real-Time Locating System to Measure Walking Activity Associated with Wandering Behaviors Among Institutionalized Older Adults

A real-time locating system (RTLS) can be used to track the walking activity of institutionalized older adults in long-term care who are at risk for wandering behaviors. The benefits of a RTLS are objective and continuous measurements of activity. Self-report methods of activity, especially wandering, by health care staff are vulnerable to floor effects and recall bias, and continuous clinical or research observation over the long-term can be time-consuming and expensive. Health care staff also fail to recognize the onset and/or duration of wandering behaviors, which are associated with a variety of adverse health outcomes in this population but amenable to intervention. RTLS technologies can measure the walking activity of institutionalized residents with cognitive impairment over time with a high degree of accuracy. This is particularly useful for the study of wandering, defined as walking for at least 60 seconds with few (if any) breaks in activity. Wandering is associated with disease progression, hospitalizations, falls and death. Previous work suggests older adults with poor balance ability and high sustained walking activity may be particularly susceptible to poor health outcomes. RTLS&#39;s are used to assess cognitive impairment and factors associated with gait and balance; however, supplemental paper and pencil gait/balance tools may be used to further refine risk profiles. This project discusses the use of a RTLS to measure walking activity and also gait quality and balance ability measures on this population.

This paper discusses the use of a continuous and objective real-time locating system to measure walking activity associated with wandering behaviors, focusing on older adults with cognitive impairment. Walking activity is measured by walking distance, sustained walking distance, and sustained gait speed. Also assessed are gait quality and balance ability.

A real-time locating system (RTLS) can be used to track the walking activity of institutionalized older adults in long-term care who are at risk for ...

Behavior

An Instrumented Pull Test to Characterize Postural Responses

Impairment of postural reflexes, termed postural instability, is a common and disabling deficit in Parkinson's disease. To assess postural reflexes, clinicians typically employ the pull test to grade corrective responses to a backward perturbation at the shoulders. However, the pull test is prone to issues with reliability and scaling (score/4). Here, we present an instrumented version of the pull test to more precisely quantify postural responses. Akin to the clinical test, pulls are manually administered except pull force is also recorded. Displacements of the trunk and feet are captured by a semi-portable motion tracking system. Raw data represent distance traveled (in millimeter units), making subsequent interpretation and analysis intuitive. The instrumented pull test also detects variabilities influencing pull test administration, such as pull force, thereby identifying and quantifying potential confounds that can be accounted for by statistical techniques. The instrumented pull test could have application in studies seeking to capture early abnormalities in postural responses, track postural instability over time, and detect responses to therapy.

Impairment of postural reflexes, termed postural instability, is difficult to quantify. Clinical assessments such as the pull test suffer issues with reliability and scaling. Here, we present an instrumented version of the pull test to objectively characterize postural responses.

Impairment of postural reflexes, termed postural instability, is a common and disabling deficit in Parkinson's disease. To assess postural reflexes, ...

A System for Tracking the Dynamics of Social Preference Behavior in Small Rodents

Exploring the neurobiological mechanisms of social behavior requires behavioral tests that can be applied to animal models in an unbiased and observer-independent manner. Since the beginning of the millennium, the three-chamber test has been widely used as a standard paradigm to evaluate sociability (social preference) and social novelty preference in small rodents. However, this test suffers from multiple limitations, including its dependence on spatial navigation and negligence of behavioral dynamics. Presented and validated here is a novel experimental system that offers an alternative to the three-chamber test, while also solving some of its caveats. The system requires a simple and affordable experimental apparatus and publicly available open-source analysis system, which automatically measures and analyzes multiple behavioral parameters at individual and population levels. It allows detailed analysis of the behavioral dynamics of small rodents during any social discrimination test. We demonstrate the efficiency of the system in analyzing the dynamics of social behavior during the social preference and social novelty preference tests as performed by adult male mice and rats. Moreover, we validate the ability of the system to reveal modified dynamics of social behavior in rodents following manipulations such as whisker trimming. Thus, the system allows for rigorous investigation of social behavior and dynamics in small rodent models and supports more accurate comparisons between strains, conditions, and treatments.

Described here is a novel automated experimental system that offers an alternative to the three-chamber test and also solves several caveats. This system supplies multiple behavioral parameters that enable rigorous analysis of small rodent behavioral dynamics during the social preference and social novelty preference tests.

Exploring the neurobiological mechanisms of social behavior requires behavioral tests that can be applied to animal models in an unbiased and ...

A Modified Lean and Release Technique to Emphasize Response Inhibition and Action Selection in Reactive Balance

Assessment of reactive balance traditionally imposes some type of perturbation to upright stance or gait followed by measurement of the resultant corrective behavior. These measures include muscle responses, limb movements, ground reaction forces, and even direct neurophysiological measures such as electroencephalography. Using this approach, researchers and clinicians can infer some basic principles regarding how the nervous system controls balance to avoid a fall. One limitation with the way in which these assessments are currently used is that they heavily emphasize reflexive actions without any need to revise automatic postural reactions. Such an exclusive focus on these highly stereotypical reactions would fail to adequately address how we can modify these reactions should the need arise (e.g., avoiding an obstacle with a recovery step). This would appear to be a glaring omission when one considers the enormous complexity of the environments we face daily. Overall, the status quo when evaluating the neural control of balance fails to truly expose how higher brain resources contribute to preventing falls in complex settings. The present protocol offers a way to require suppression of automatic, but inappropriate corrective balance reactions, and force a selection among alternative action choices to successfully recover balance following postural perturbation.

Here we offer a protocol that allows the user to selectively change affordances and/or constraints on movements that are relevant for recovering balance after postural perturbation.

Assessment of reactive balance traditionally imposes some type of perturbation to upright stance or gait followed by measurement of the resultant ...

Microfluidic Platform with Multiplexed Electronic Detection for Spatial Tracking of Particles

Microfluidic processing of biological samples typically involves differential manipulations of suspended particles under various force fields in order to spatially fractionate the sample based on a biological property of interest. For the resultant spatial distribution to be used as the assay readout, microfluidic devices are often subjected to microscopic analysis requiring complex instrumentation with higher cost and reduced portability. To address this limitation, we have developed an integrated electronic sensing technology for multiplexed detection of particles at different locations on a microfluidic chip. Our technology, called Microfluidic CODES, combines Resistive Pulse Sensing with Code Division Multiple Access to compress 2D spatial information into a 1D electrical signal. In this paper, we present a practical demonstration of the Microfluidic CODES technology to detect and size cultured cancer cells distributed over multiple microfluidic channels. As validated by the high-speed microscopy, our technology can accurately analyze dense cell populations all electronically without the need for an external instrument. As such, the Microfluidic CODES can potentially enable low-cost integrated lab-on-a-chip devices that are well suited for the point-of-care testing of biological samples.

We demonstrate a microfluidic platform with an integrated surface electrode network that combines resistive pulse sensing (RPS) with code division multiple access (CDMA), to multiplex detection and sizing of particles in multiple microfluidic channels.

Microfluidic processing of biological samples typically involves differential manipulations of suspended particles under various force fields in order to ...

Design and Application of a Fault Detection Method Based on Adaptive Filters and Rotational Speed Estimation for an Electro-Hydrostatic Actuator

The electro-hydrostatic actuator (EHA) is a promising actuating apparatus used in flight control systems for more electric aircraft (MEA) due to its high power density and low maintenance. Since the reliability of the system decreases with increasing complexity, fault detection is becoming increasingly important. In this paper, an adaptive filter was designed based on a normalized least mean square (NLMS) algorithm, which could identify the resistance of the motor windings online to detect electrical faults in the EHA. Additionally, based on the analytical relationship between rotational speed and displacement, a rotational speed estimation method was designed. By comparing the actual rotational speed with the estimated one, hydraulic faults could be detected. To verify the efficacy of the aforementioned method, software was applied for the modeling and simulations, which included fault injection and detection. On this basis, an experimental platform was built and then subjected to a series of validation experiments. The results indicate that the fault detection method has the potential to detect electrical and hydraulic faults in an EHA.

In this paper, an adaptive filter based on a normalized least mean square (NLMS) algorithm and a rotational speed estimation method are introduced to detect the electrical and hydraulic faults of the electro-hydrostatic actuator (EHA). The efficacy and feasibility of the aforementioned methods are verified through simulations and experiments.

The electro-hydrostatic actuator (EHA) is a promising actuating apparatus used in flight control systems for more electric aircraft (MEA) due to its high ...

Enhancing Classroom Engagement: AI-Based Analysis of Student Attention Levels

Artificial Intelligence-Based System for Detecting Attention Levels in Students

The attention level of students in a classroom can be improved through the use of Artificial Intelligence (AI) techniques. By automatically identifying the attention level, teachers can employ strategies to regain students' focus. This can be achieved through various sources of information.
One source is to analyze the emotions reflected on students' faces. AI can detect emotions, such as neutral, disgust, surprise, sadness, fear, happiness, and anger. Additionally, the direction of the students' gaze can also potentially indicate their level of attention. Another source is to observe the students' body posture. By using cameras and deep learning techniques, posture can be analyzed to determine the level of attention. For example, students who are slouching or resting their heads on their desks may have a lower level of attention. Smartwatches distributed to the students can provide biometric and other data, including heart rate and inertial measurements, which can also be used as indicators of attention. By combining these sources of information, an AI system can be trained to identify the level of attention in the classroom. However, integrating the different types of data poses a challenge that requires creating a labeled dataset. Expert input and existing studies are consulted for accurate labeling. In this paper, we propose the integration of such measurements and the creation of a dataset and a potential attention classifier. To provide feedback to the teacher, we explore various methods, such as smartwatches or direct computers. Once the teacher becomes aware of attention issues, they can adjust their teaching approach to re-engage and motivate the students. In summary, AI techniques can automatically identify the students' attention level by analyzing their emotions, gaze direction, body posture, and biometric data. This information can assist teachers in optimizing the teaching-learning process.

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

This paper proposes an artificial intelligence-based system to automatically detect whether students are paying attention to the class or are distracted. This system is designed to help teachers maintain students' attention, optimize their lessons, and dynamically introduce modifications in order for them to be more engaging.

The attention level of students in a classroom can be improved through the use of Artificial Intelligence (AI) techniques. By automatically identifying ...

An Experimental Approach to Induce Trips in Lower-Limb Amputees

Reestablishing balance after a trip is challenging for lower-limb amputees and often results in a fall. The effectiveness of reestablishing balance following a trip depends on factors such as amputation level (transtibial or transfemoral) or which limb is tripped (prosthetic or sound/lead or trailing). Understanding the recovery responses can help identify strategies to avoid a trip becoming a fall and what trip-response functionality could be designed into a prosthesis. This study presents an experimental approach for inducing unexpected trips in individuals with amputation. Tripping was manually triggered by activating an electromagnetic device to raisea polypropylene wire to obstruct (bring to a near halt) theswinging limb during its mid-swing phase. A safety harness attached to a ceiling rail ensured participants did not hit the ground if they failed to reestablish balance following the trip (i.e., it prevented a fall from occurring). One transtibial amputee completed repeated walking trials in which a trip was induced around 1 out of 15 times to avoid it being anticipated. 3D kinematics were determined via two smartphones (60Hz) using the OpenCap software, highlighting that the experimental approach induced meaningful tripping/recovery responses dependent on which limb was tripped (prosthetic or sound). The presented methodology avoids using a rigid obstacle, potentially reducing the risk of injuries, and is inexpensive and easy to set up. Importantly it permits a trip to be unexpectedly introduced during the mid-swing phase of the gait and hence provides an approach for identifying real-world trip recovery responses. When tripping the sound limb, participants could &#39;disentangle&#39; from the trip-wire (post-trip) by plantarflexing the ankle, but such action was not possible when tripping the prosthetic limb.

An experimental approach was developed to induce trips in lower-limb amputees. The goal was to create unexpected trips and induce meaningful tripping/recovery responses. The kinematic data from one transtibial amputee confirmed that such an approach effectively elicits reactive trip recovery responses.

Reestablishing balance after a trip is challenging for lower-limb amputees and often results in a fall. The effectiveness of reestablishing balance ...

Design and Analysis for Fall Detection System Simplification

Samenvatting

Meer video's verkennen