Adaptive Path Exploration for Mining Truck Route Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing path planning methods for mining trucks face challenges in complex and dynamic mining environments due to fixed exploration parameters, leading to low search efficiency, poor path quality, and high maintenance costs, especially in scenarios with irregular terrain and debris.
Innovation Solution
A deep reinforcement learning-based path exploration parameter optimization system that adapts exploration parameters using a deep reinforcement learning network, incorporating environmental obstacle information to enhance search efficiency and path quality, reducing reliance on maps and overfitting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed exploration parameters are used in path planning algorithms, then the algorithm structure remains simple and easy to implement, but the search efficiency and path quality deteriorate in complex and dynamic mining environments
Solution Approach 1:
The patent applies dynamics by transitioning from fixed exploration parameters to adaptive parameters that change based on environmental conditions. The deep reinforcement learning network dynamically adjusts path exploration parameters according to the current state of the mining environment, enabling the algorithm to adapt to complex and dynamic scenarios while maintaining reasonable structural complexity.
Solution Approach 2:
The patent implements parameter changes by using a deep reinforcement learning network to optimize path exploration parameters based on environmental information. The network learns to adjust parameters such as step length, steering angle, and search radius dynamically, transforming the algorithm from using fixed parameters to using learned adaptive parameters that improve search efficiency in varying mining conditions.
2Adaptability or versatility
If manual rules are used to adjust exploration parameters, then some adaptability to different scenarios is achieved, but the maintenance cost and parameter tuning complexity increase significantly
Solution Approach 1:
The patent applies self-service by enabling the path planning system to automatically adjust its own parameters through the deep reinforcement learning network. The network learns from environmental interactions and autonomously optimizes exploration parameters without requiring manual intervention or extensive parameter tuning, reducing maintenance complexity while maintaining high adaptability to different mining scenarios.
Solution Approach 2:
The patent implements feedback by using the deep reinforcement learning network to continuously learn from environmental outcomes and adjust parameters accordingly. The network receives feedback from the environment about the effectiveness of parameter choices and uses this information to improve future parameter selections, creating a self-optimizing system that adapts to scenarios without manual tuning.
3Speed
If sampling-based methods like rapidly-exploring random tree algorithm are used, then computation speed is improved, but the path quality and conformability to motion characteristics deteriorate
Solution Approach 1:
The patent applies parameter changes by using the deep reinforcement learning network to optimize exploration parameters such as step length and steering angle based on environmental conditions. This allows the system to maintain faster computation speeds while improving path quality by adapting parameters to better conform to the motion characteristics of mining trucks in different scenarios.
Data Source
AI summary
The present invention relates to the technical field of path planning, and provides a deep reinforcement learning-based path exploration parameter optimization system. The system comprises: a variable parameter path planning module, configured to perform node exploration based on a deep reinforcement learning network, conduct collision detection on child nodes in a child node set, calculate cost values for all child nodes, and finally generate a loading and parking path using a Reeds-Shepp curve; an environmental state space modeling module, configured to perform regional division of obstacles around a current node and conduct environmental state space modeling; and a deep learning parameter optimization module, configured to construct a deep learning network to compute an optimal step size and an optimal steering angle, build a reward function to optimize the deep learning network, and simultaneously execute a training process of the deep learning network.