Abstract Human activity recognition (HAR) has recently gained popularity due to its applications in healthcare, surveillance, human‐robot interaction, and various other fields. Deep learning (DL)‐based models have been successfully applied to the raw data captured through inertial measurement unit (IMU) sensors to recognize multiple human activities. Despite the success of DL‐based models in human activity recognition, feature extraction remains challenging due to class imbalance and noisy data. Additionally, selecting optimal hyperparameter values for DL models is essential since they affect model performance. The hyperparameter values of some of the existing DL‐based HAR models are chosen randomly or through the trial‐and‐error method. The random selection of these significant hyperparameters may be suitable for some applications, but sometimes it may worsen the model's performance in others. Hence, to address the above‐mentioned issues, this research aims to develop an optimized DL model capable of recognizing various human activities captured through IMU sensors. The proposed DL‐based HAR model combines convolutional neural network (CNN) layers and bidirectional long short‐term memory (Bi‐LSTM) units to simultaneously extract spatial and temporal sequence features from raw sensor data. The Rao‐3 metaheuristic optimization algorithm has been adopted to identify the ideal hyperparameter values for the proposed DL model in order to enhance its recognition performance. The proposed DL model's performance is validated on PAMAP2, UCI‐HAR, and MHEALTH datasets and achieved 94.91%, 97.16%, and 99.25% accuracies, respectively. The results reveal that the proposed DL model performs better than the existing state‐of‐the‐art (SoTA) models.

Abstract The research paper presents the heterogeneous computing model for analysis & restoration of human walking deformity and posture instability. Gait‐related walking activities are very important for the analysis of postural instability, repairment of gait abnormality, diagnosis of cognitive declination, enhance the cognitive ability of human‐centered humanoid robot system, and many clinical diagnoses, for example, Parkinson, pathological gait, freezing of gait, etc. at an early stage. For experiment analysis, 10 different lower limb activities are being considered of healthy and crouch walking subjects. A total of 25 healthy and 10 crouch walk subjects are considered for experiment purposes of different age groups, sex, and mental status. To achieve this objective the pattern of 10 different rehabilitation activities are captured using RGB‐Depth (RGB‐D) camera and classified using heterogeneous deep learning models. Different deep learning models Convolutional Neural Network (CNN) and CNN‐LSTM (CNN‐Long Short Term Memory) are used for the classification of these rehabilitation exercises. The RGB‐D data is obtained using a Microsoft Kinect v2 sensor on a 100 Hz sampling frequency. Experimental results have shown significant activity recognition accuracy with 96% and 98% for CNN and CNN‐LSTM models respectively.

Abstract Human walking is a very challenging task and always requires rigorous practice. It is a learning process that involves the complex coordination of the brain and lower limbs. The bipedal robots that mimic the human morphological structure to produce human similar walking, are not capable of producing an efficient walk. Due to walking challenges and structural differences, a robot cannot walk like a human being. In this research, to achieve the aforementioned objective to produce a human similar walk, human lower extremity activities are considered to understand walking behaviour. The experiment involves different walking styles on different terrains. To capture the learning process of bipedal robot locomotion, a deep learning‐based ensemble classifier is introduced for human lower activities recognition. To understand the learning process seven different walking activities are considered for analysis purposes. An Inertial measurement unit (IMU) is used as a wearable device due to its small form factor and unobtrusive nature to capture the walking movement of different lower limbs joints. Three public datasets viz. mHealth, OU‐ISIR similar action and HAPT inertial sensor data sets are considered for this study. To classify the activities, 2 different deep learning models namely convolutional neural network (CNN) and long short‐term memory (LSTM) are used. To generalize the results, an ensemble of different classifiers is implemented. The Classifier has reported accuracy of 99.25%, 88.48% and 97.44%, respectively, on the aforementioned data sets. This work can be utilized for elderly subjects' postural stability, rehabilitation of patients post‐stroke and trauma, generation of robot walk trajectories in cluttered environment and reconstruction of impaired walking.

In this paper, we have designed the vector fields (VFs) for all the six joints (hip, knee, and ankle) of a bipedal walking model. The bipedal gait is the manifestation of temporal changes in the six joints angles, two each for hip, knee, and ankle values and it is a combination of seven different discrete subphases. Developing the correct joint trajectories for all the six joints was difficult from a purely mechanics-based model due to its inherent complexities. To get the correct and exact joint trajectories, it is very essential for a modern bipedal robot to walk stably. By designing the VF correctly, we are able to get the stable joint trajectory ranges and able to reproduce angle ranges from theses designed VFs. This is purely a data driven computational modeling approach, which is based on the hypothesis that morphologically similar structure (human-robot) can adopt similar gait patterns. To validate the correctness of the design, we have applied all the possible combination of joint trajectories to HOAP-2 bipedal robot, which could walk successfully maintaining its stability. The VF provides joint trajectories for a particular joint. The results show that our data driven computational model is able to provide the correct joints angle ranges, which are stable.