Obstacle avoidance method for robot, obstacle avoidance device for robot, and non-transitory computer readable storage medium

The robot performs preliminary exploration to identify trap regions before actual cleaning operations. During exploration, it detects and records locations where it becomes trapped, storing these as trap region information. When subsequent cleaning operations encounter these pre-identified trap regions, the robot can proactively avoid them by selecting alternative paths, thus resolving the contradiction between reliable obstacle avoidance and adaptability to diverse trap scenarios.

The system pre-processes environmental information by identifying and marking trap regions during initial exploration phases. This preliminary identification allows the robot to build a knowledge base of problematic areas before actual cleaning tasks begin, enabling more reliable and adaptable obstacle avoidance during subsequent operations without requiring complex real-time decision-making for each new scenario.

The robot implements periodic obstacle detection rather than continuous monitoring. During normal travel, it performs detection at regular intervals or at key decision points. When approaching previously identified trap regions or encountering uncertain situations, it increases detection frequency. This periodic approach maintains reliable real-time obstacle detection capability while significantly reducing overall energy consumption compared to continuous monitoring.

The detection system dynamically adjusts its operation mode based on the robot's current state and environmental context. It switches between active detection, passive monitoring, and prediction modes. For example, when traveling through known safe areas, it uses passive monitoring or prediction based on stored trap region data; when approaching uncertain areas or trap regions, it activates active detection. This dynamic adjustment maintains detection reliability while optimizing energy consumption.

The system incorporates feedback mechanisms where the robot's past experiences with trap regions are stored and fed back into its path planning system. When the robot encounters a trap region during cleaning operations, it records this information and uses it to adjust future route selections. The feedback loop continuously improves the robot's ability to avoid previously problematic areas while maintaining efficient cleaning coverage, thus resolving the contradiction between productivity and operational continuity.

The cleaning route is dynamically adjusted based on real-time detection and historical trap region data. Rather than following rigid preset routes, the system modifies paths on-the-fly to avoid identified trap regions. When a trap region is detected or predicted along the current path, the robot dynamically recalculates alternative routes that maintain cleaning efficiency while avoiding repetitive trapping, thus balancing productivity and operational continuity.