Mobile Robot Path Planning Using Bayesian Network Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional path planning methods for mobile robots focus solely on the shortest distance, neglecting environmental characteristics and human activity, leading to potential collisions and inefficiencies in navigation, especially in environments with frequent human movement or sudden obstacles.
Innovation Solution
A path planning method that utilizes real-time environmental data, including grid maps, obstacle maps, and Bayesian networks to create a safe and efficient navigation path by detouring around areas with high human activity or potential obstacles, incorporating congestion maps to optimize route planning and avoid collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the robot moves along the shortest path from current position to destination point, then the movement distance is minimized, but the robot has a high possibility of colliding with humans or moving obstacles
Solution Approach 1:
The system performs preliminary learning of human movement patterns and environmental characteristics before actual navigation. The Bayesian network is pre-trained with historical data about human activities and obstacle locations, enabling the robot to predict dangerous areas in advance and plan safe paths proactively rather than reactively avoiding collisions
Solution Approach 2:
The path planning system dynamically adapts to changing environmental conditions by continuously updating the Bayesian network with real-time sensor data. The congestion map and human activity patterns are refreshed based on current observations, allowing the robot to adjust its navigation strategy dynamically rather than following static pre-defined paths
2Reliability
If the robot detours around areas with high human activity to avoid collisions, then collision avoidance is improved, but the movement distance and time increase
Solution Approach 1:
The system applies different path planning strategies to different spatial regions based on local characteristics. High-probability dangerous areas identified by the Bayesian network receive special attention with detour paths, while low-risk areas allow for more direct routing. This localized approach ensures safety where needed while maintaining efficiency in safe zones
Solution Approach 2:
The system changes the cost parameters in path planning based on predicted danger levels. Areas with high probability of human presence or obstacles are assigned higher cost values in the navigation function, dynamically adjusting the path optimization criteria rather than using fixed uniform parameters throughout the environment
3Reliability
If the robot frequently changes movement direction to avoid sudden obstacles, then collision avoidance is improved, but the navigation speed and efficiency decrease
Solution Approach 1:
The Bayesian network predicts potential obstacle locations and human movement trajectories in advance, allowing the robot to plan smooth curved paths around anticipated dangerous zones rather than making abrupt reactive direction changes. This preliminary prediction enables maintainance of higher speeds through pre-computed safe trajectories
Data Source
AI summary
The present invention provides a safe path planning method including a process of using a grid map and a congestion map, and a process of planning a path of a mobile robot using a cost function.