Bidirectional Long-Short Term Memory Based Approach for Sampling in Dual Arm Locomotion

Bimanual movement involves coordinated movement of both limbs, a multifaceted challenge in robotics and biomechanics. The efficiency of movement planning in bimanual movement is of utmost importance in many applications, from developing humanoid robots to designing rehabilitation devices. This paper presents an innovative method that uses Bi-LSTM (bidirectional long-short-term memory) networks to improve motion sampling strategies for bimanual movements. The proposed approach uses the temporal dependencies and bidirectional contextual information provided by Bi-LSTM networks to generate adaptive and naturalistic movement trajectories, effectively capturing the complex nuances of interlimb coordination. Results highlight the superior performance of our method over existing techniques. This study provides a promising solution to the complexities of movement planning for bimanual movement, which has wide applications in robotics, assistive devices, and healthcare. In addition, it is strategically aligned with an emerging trend in the manufacturing industry - the integration of collaborative robots (cobots) with 5 and 7 degrees of freedom - which improves production flexibility and agility. At the same time, it reiterates the core principles of Industry 4.0 and emphasizes the development of intelligent, interconnected production systems that use advanced robotics and automation to increase work efficiency and adaptability in the manufacturing environment.