Citation: Son, C.-S.; Kang, W.-S. Multivariate CNN Model for Human Locomotion Activity Recognition with a Wearable Exoskeleton Robot. Bioengineering 2023, 10, 1082. https://doi.org/10.3390/ bioengineering10091082 Academic Editors: Crescenzio Gallo and Gianluca Zaza Received: 11 August 2023 Revised: 5 September 2023 Accepted: 9 September 2023 Published: 13 September 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). bioengineering Article Multivariate CNN Model for Human Locomotion Activity Recognition with a Wearable Exoskeleton Robot Chang-Sik Son 1 and Won-Seok Kang 1,2, * 1 Division of Intelligent Robot, Daegu Gyeongbuk Institute of Science & Technology (DGIST), Daegu 42988, Republic of Korea; changsikson@dgist.ac.kr 2 Department of Biomedical Science, Graduate School, Kyungpook National University, Daegu 41944, Republic of Korea * Correspondence: wskang@dgist.ac.kr Abstract: This study introduces a novel convolutional neural network (CNN) architecture, encom- passing both single and multi-head designs, developed to identify a user’s locomotion activity while using a wearable lower limb robot. Our research involved 500 healthy adult participants in an activi- ties of daily living (ADL) space, conducted from 1 September to 30 November 2022. We collected prospective data to identify five locomotion activities (level ground walking, stair ascent/descent, and ramp ascent/descent) across three terrains: flat ground, staircase, and ramp. To evaluate the predictive capabilities of the proposed CNN architectures, we compared its performance with three other models: one CNN and two hybrid models (CNN-LSTM and LSTM-CNN). Experiments were conducted using multivariate signals of various types obtained from electromyograms (EMGs) and the wearable robot. Our results reveal that the deeper CNN architecture significantly surpasses the performance of the three competing models. The proposed model, leveraging encoder data such as hip angles and velocities, along with postural signals such as roll, pitch, and yaw from the wearable lower limb robot, achieved superior performance with an inference speed of 1.14 s. Specifically, the F-measure performance of the proposed model reached 96.17%, compared to 90.68% for DDLMI, 94.41% for DeepConvLSTM, and 95.57% for LSTM-CNN, respectively. Keywords: human activity recognition; wearable robot; single-head CNN; multi-head CNN; hyperparameter optimization; time series classification 1. Introduction Wearable exoskeleton robots have been developed to aid individuals in a range of activities, including carrying heavy objects, alleviating the burden of physically demanding tasks, and assisting in-patient rehabilitation. Studies have indicated that exoskeletons can substantially assist and lower metabolic costs during walking [1,2]. Numerous powered exoskeleton robots have facilitated the improvement of lower extremity movement deficits resulting from strokes [3–5] or injuries such as amputations [6,7] by applying assistive torques to the joints [8]. However, despite these successful applications, several challenges persist in developing safe and versatile control systems [9], including the identification of the wearer’s intended movement without external commands, and the autonomous transition between different activity-specific controllers. One approach to identifying intended activity involves using a locomotor activity intent recognition framework [10,11]. This method is predominantly applied in medi- cal rehabilitation, analyzing patients’ gait patterns to furnish clinicians with a quantita- tive overview of motor function behavior over extended durations, thus aiding objective treatment strategy applications [12]. For instance, due to postural instability and gait disturbances, Parkinson’s disease patients have an increased susceptibility to fall-related injuries [13,14]. Real-time movement monitoring can mitigate injury risks by promptly Bioengineering 2023, 10, 1082. https://doi.org/10.3390/bioengineering10091082 https://www.mdpi.com/journal/bioengineering