Falls are a major health issue in societies globally and the second leading cause of unintentional death worldwide. To address this issue, many studies aim to remotely monitor gait to prevent falls. However, these activity data collected in studies must be labelled with the appropriate environmental context through Human Activity Recognition (HAR). Multimodal HAR datasets often achieve high accuracies at the cost of cumbersome sensor systems, creating a need for these datasets to be analysed to identify the sensor types and locations that enable high-accuracy HAR. This paper analyses four datasets, USC-HAD, HuGaDB, Camargo et al.’s dataset, and CSL-SHARE, to find optimal models, methods, and sensors across multiple datasets. Regarding window size, optimal windows are found to be dependent on the sensor modality of a dataset but mostly occur in the 2–5 s range. Support Vector Machines (SVMs) and Artificial Neural Networks (ANNs) are found to be the highest-performing models overall. ANNs are further used to create models trained on the features from individual sensors of each dataset. From this analysis, Inertial Measurement Units (IMUs) and three-axis goniometers are shown to be individually capable of high classification accuracy, with Electromyography (EMG) sensors exhibiting inconsistent and reduced accuracies. Finally, it is shown that the thigh is the optimal location for IMU sensors, with accuracy decreasing as IMUs are placed further down away from the thigh.

Falls are a prominent issue in society and the second leading cause of unintentional death globally. Traditional gait analysis is a process that can aid in identifying factors that increase a person’s risk of falling through determining their gait parameters in a controlled environment. Advances in wearable sensor technology and analytical methods such as deep learning can enable remote gait analysis, increasing the quality of the collected data, standardizing the process between centers, and automating aspects of the analysis. Real-world gait analysis requires two problems to be solved: high-accuracy Human Activity Recognition (HAR) and high-accuracy terrain classification. High accuracy HAR has been achieved through the application of powerful novel classification techniques to various HAR datasets; however, terrain classification cannot be approached in this way due to a lack of suitable datasets. In this study, we present the Context-Aware Human Activity Recognition (CAHAR) dataset: the first activity- and terrain-labeled dataset that targets a full range of indoor and outdoor terrains, along with the common gait activities associated with them. Data were captured using Inertial Measurement Units (IMUs), Force-Sensing Resistor (FSR) insoles, color sensors, and LiDARs from 20 healthy participants. With this dataset, researchers can develop new classification models that are capable of both HAR and terrain identification to progress the capabilities of wearable sensors towards remote gait analysis.