Application of autonomous pathfinding system to kinematics and dynamics problems by implementing network constraints

TL;DR

This paper introduces a neural network-based model for robotic arm kinematics and dynamics that incorporates network constraints, enabling adaptive solution finding and robustness to perturbations, with potential applications in robotics and parallel computation.

Contribution

The paper presents a novel network model that integrates physical constraints and obstacles, maintaining a consistent framework as constraints increase, and supports adaptive, distributed problem solving.

Findings

The model successfully finds solutions for robot arm kinematics and dynamics.

It automatically adapts to perturbations like obstacle presence or target shifts.

The framework is scalable with increasing constraints and suitable for parallel algorithms.

Abstract

A neural network system in an animal brain contains many modules and generates adaptive behavior by integrating the outputs from the modules. The mathematical modeling of such large systems to elucidate the mechanism of rapidly finding solutions is vital to develop control methods for robotics and distributed computation algorithms. In this article, we present a network model to solve kinematics and dynamics problems for robot arm manipulation. This model represents the solution as an attractor in the phase space and also finds a new solution automatically when perturbations such as variations in the end position of the arm or obstacles occur. In the proposed model, the physical constraints, target position, and the existence of obstacles are represented by network connections. Therefore, the theoretical framework of the model remains almost the same when the number of constraints increases. In addition, as the model is regarded as a distributed system, it can be applied toward the development of parallel computation algorithms.