Robot Trap Region Memory for Autonomous Obstacle Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Sweeping robots often get trapped at obstacles such as furniture or small objects, requiring user intervention to continue working, leading to poor user experience.
Innovation Solution
The robot uses sensors and image recognition to identify obstacles, labels their locations, and avoids them in subsequent travels, using a control system to adjust its route and movement to prevent re-trapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional obstacle avoidance methods are used, then the robot can navigate environments, but the planning time is long and real-time response is poor
Solution Approach 1:
The patent segments the obstacle avoidance planning into two distinct phases: global path planning (using A* algorithm for overall route) and local path planning (using dynamic window approach for real-time adjustments). This segmentation allows each phase to operate independently with appropriate algorithms, improving overall response speed while reducing total planning time.
Solution Approach 2:
The patent performs preliminary global path planning using the A* algorithm to establish a baseline route before executing local real-time adjustments. This preliminary action provides a pre-computed path framework that reduces the computational burden during real-time local planning, thereby decreasing overall planning time while maintaining responsiveness.
2Measurement precision
If complex path planning algorithms are used, then the path planning precision is improved, but the computational complexity increases
Solution Approach 1:
The patent divides the path planning into global planning (A* algorithm for precise route calculation) and local planning (dynamic window approach for real-time adjustments). This segmentation assigns different computational complexities to different phases, achieving high precision where needed while managing overall computational load.
Solution Approach 2:
The patent applies the A* algorithm globally to compute a precise baseline path, then uses simpler local adjustments for real-time modifications. This partial application of complex algorithms only where necessary (global planning) while using simpler methods for routine adjustments optimizes the balance between precision and computational complexity.
3Adaptability or versatility
If the robot follows a pre-planned path strictly, then the path following is simple, but the adaptability to dynamic obstacles is poor
Solution Approach 1:
The patent implements a two-layer planning system where the global path provides a static framework and the local planner dynamically adjusts the trajectory in real-time based on sensor feedback. This dynamic approach enables the robot to adapt to moving obstacles while maintaining a structured control architecture that manages complexity.
Solution Approach 2:
The patent incorporates real-time sensor feedback into the local path planning process, allowing the robot to detect and respond to dynamic obstacles. The feedback loop continuously monitors the environment and adjusts the path accordingly, improving adaptability while using the hierarchical structure to manage control complexity.
4Measurement precision
If multiple sensors are deployed for obstacle detection, then the detection precision is improved, but the system complexity increases
Solution Approach 1:
The patent combines data from multiple sensors (laser radar, ultrasonic sensors, infrared sensors) into a unified obstacle detection and avoidance system. By merging sensor inputs and processing them through a coordinated control architecture, the system achieves high detection precision while managing overall system complexity through integration.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
A robot obstacle avoidance method, a device, and a storage medium. The method comprises: acquiring, in real-time, trapping feature information during a traveling process, the trapping feature information comprising a sensor parameter and/or image information (step S502); if the trapping feature information meets a trapping condition, checking whether a current position is within a trap region, wherein the trap region refers to a region in which a robot was trapped before or is likely to become trapped in (step S504); and if so, terminating a current movement route, and moving out of the trap region (S506). The method enables a robot to automatically acquire, according to its own sensor, an operation state of the robot during a traveling process, to record a current position when its own sensor determines that the robot is trapped, and to automatically detour and avoid the position if the robot is operating near the position again in the future. In this way, the invention enables continuous operation of the robot, and enhances smoothness of automatic cleaning by the robot.