Multi-robotic systems of varying sizes and levels of complexity can effortlessly address real-world distributed tasks like exploration, catastrophic surveillance, logistics, and industrial manufacturing. To perform such tasks efficiently, individual robots should communicate and collaborate with each other to cooperatively map the area of interest. The primary advantage of multi-robot systems over a single stand-alone robot is the enhanced spatial coverage, significantly reducing the exploration time and energy required for unknown terrain mapping. However, a large number of robots can increase the overall cost as well as the complexity of the operation by introducing additional redundancy. Therefore, the objective of this research is to study the functionality of multi-robot systems by investigating map-merging and nearest-robot navigation using a case study of Turtlebot3. In this study, computational simulations are performed in the ROS framework where multiple robots are deployed to generate individual local maps which are further transformed to create a global world map. This global map is used to compute the Euclidean distance subsequently enabling the nearest member goal navigation. Indoor experimental tests are performed to verify the simulation study, demonstrating that map merging and localised nearest member navigation can cut down operational time. The overall operational time was reduced by 23.43 % with the deployment of two robots and 45.85 % with three robots in the simulation environment, hence verifying the reduction in mission completion time.

Squat is a closed kinetic chain exercise widely used for strength and conditioning applications. This exercise also supports the preparedness for multiple sports. In Physical rehabilitation, squatting is widely incorporated to strengthen the disorders and activate the muscle extensors. This study collected 3D Motion capture data from 5 healthy, trained individuals. Subjects with a history of injury were excluded. The study aims to examine the difference in kinematic variables between Front and Back squats without load conditions and their postural changes. The Kinematic profile was analyzed using computational musculoskeletal software - OpenSim. Based on the statistical results, it is indicated that there is no significant difference in the lower body joints (p< 0.05), contradicting the upper body joints (p >0.05), especially the spine. The peak joint angle value for the neck, spine, and shoulder in the sagittal plane during the Front squat were -21.53, 16.22, and 22.31 degrees, and during the Back squat were -31.68, 9.19, and 70.90 degrees, respectively. Further statistical test was performed using the Paired sample t-test, indicating a significant difference in knee, neck, spine, and shoulder. Hence, this study differentiates the joint angle variation during the two squat styles. The existence of a dominant component in the squat technique has been identified which potentially adds value and helps in curating the needs of patients requiring rehabilitative techniques. The variables analyzed in this study help identify additional parameters that aid in the qualitative and quantitative analysis of a dedicated posture restoration scheme, including squatting as part of the prescribed exercises.

Neurodegenerative diseases (NDDs) such as Parkinson’s disease (PD) and stroke often manifest as altered gait rhythms. While various clinical and progression metrics are available to evaluate gait dynamics, the underlying sensorimotor mapping between normal and abnormal gait patterns remains insufficiently explored. These gait patterns are inherently non-stationary, non-linear, and non-parametric, posing significant challenges in their characterisation. Spatio-temporal to spatio-spectral features were essential for capturing the intricate complexities of impaired gait dynamics, as reflected in fractal dimensional components and entropy-based nonlinear feature space mappings. This systematic review chapter aims to bridge the existing knowledge gaps by providing comprehensive insights into the sensorimotor alterations observed in NDDs. Recent feature engineering and gait analysis advancements offer a detailed understanding of how frequency-domain and spectral-level insights can enhance disease detection and monitoring. Additionally, the findings presented here contribute to the development of non-invasive neuromodulation techniques, which hold promise in restoring motor function and improving the quality of life of patients. Overall, this review is a valuable resource for clinicians, researchers, and engineers seeking to explore the sensorimotor mechanisms underlying gait abnormalities further. It also underscores the importance of applying robust machine learning (ML) algorithms and feature selection approaches to achieve more accurate and reliable assessments of NDD progression.

Visual observations and kinematic ratings derived from inertial devices are still common in gait assessment; however, they frequently lack clinical specificity across diverse stroke diseases. The neural drive signals necessary for reliable functional ambulation classification are often ignored by kinematic-based classifiers alone. To address this, we propose a multi-modal feature integration framework that combines joint angles, centre of mass, and temporal components of muscle synergies. The proposed approach primarily aims to improve classification accuracy while enhancing clinical interpretability by linking impairment severity to motor primitives. This framework provides a data-informed foundation for more personalised assessment and targeted rehabilitation planning. Based on the multi-modal fusion of the feature space, the BiLSTM network is calibrated using an extensive grid search method to provide an automated, accurate assessment of the FAC score, achieving 93.10% accuracy and 0.9289 Macro F1 score with Leave-One-Subject-Out cross-validation (LOSO-CV) and outperforming the decision-making ability when kinematics, muscle synergy, or center of mass (CoM) dynamics are presented individually.