Mobile Robot Path Planning With Evolutionary Recurrent Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing path planning methods for autonomous mobile robots often fall into local optimal solutions due to the limitations of gradient descent algorithms, making it difficult to obtain the global optimal path in complex and dynamic environments.
Innovation Solution
A neural network training method based on evolutional algorithms is employed, which involves constructing recurrent neural networks, initializing and optimizing them using evolutional algorithms, and selecting the global optimal individual to obtain the global optimal neutral networks for path planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If gradient descent algorithms are used to train recurrent neural networks for path planning, then the training process is simple and computationally efficient, but the network easily falls into local optimal solutions and cannot quickly converge to the global optimal solution
Solution Approach 1:
The patent replaces the traditional gradient descent algorithm (a mechanical optimization process) with an evolutional algorithm that uses biological inspiration principles. The evolutional algorithm includes selection, crossover, and mutation operations that simulate natural evolution, allowing the neural network to escape local optima and converge to global optimal solutions more effectively without being constrained by the gradual step-by-step nature of gradient descent.
Solution Approach 2:
The patent changes the optimization parameters and methodology by introducing evolutional algorithms with adjustable parameters such as population size, mutation rate, and crossover probability. These parameter changes enable the system to dynamically adapt the search process, balancing exploration and exploitation to achieve both high optimality and efficient convergence in path planning.
2Adaptability or versatility
If recurrent neural networks with internal state feedback are used to describe nonlinear dynamic behaviors, then the dynamic approximation capability is very strong, but the network structure and training complexity increase
Solution Approach 1:
The patent replaces the complex training process of recurrent neural networks with evolutional algorithms. Instead of using traditional backpropagation through time that requires careful initialization and hyperparameter tuning, the evolutional algorithm directly optimizes the network weights and structures through selection, crossover, and mutation, simplifying the training complexity while maintaining the network's dynamic approximation capabilities.
3Measurement precision
If evolutional algorithms are used to optimize recurrent neural networks for path planning, then the global optimal solution can be obtained, but the computational complexity and training time increase
Solution Approach 1:
The patent segments the path planning problem into multiple generations of evolutional optimization, where each generation focuses on improving the solution quality through selective pressure. By dividing the optimization process into discrete generations with selection, crossover, and mutation operations, the algorithm manages computational complexity while systematically converging to global optimal solutions.
Solution Approach 2:
The patent introduces dynamic adjustments in the evolutional algorithm parameters such as mutation rate and population size during the training process. This dynamic adaptation allows the algorithm to intensify the search in promising regions and explore new areas when needed, balancing computational complexity with solution quality and convergence speed.
Data Source
AI summary
The invention discloses an adaptive path planning method based on neutral networks trained by the evolutional algorithms, the neutral network training method comprises input and output of the data acquired by the mobile sensors installed on the mobile robots as the neutral networks, and training and optimization of the recurrent neutral networks based on the evolutional algorithms; the path planning method refers to the application of the trained neutral networks to the path planning of the mobile robot, the invention effectively improves local quick search capability and global search capability of the algorithms by applying the evolutional algorithms to the optimization of the recurrent neutral networks, so that the robot can plan a rational path in a dense and uncertain environment.