A Bio-inspired Collision Avoidance Model Based on Spatial Information Derived from Motion Detectors Leads to Common...

Area of Science:

• Robotics and Artificial Intelligence
• Neuroscience and Biology

Background:

• Mobile agents require robust collision avoidance for navigation.
• Insects like flies and bees use optic flow and saccadic control for flight and gaze stabilization.
• Optic flow during translation contains depth information entangled with self-motion cues.

Purpose of the Study:

• To propose and test a computational model for collision avoidance inspired by insect behavior.
• To investigate the extraction of depth structure from optic flow using a spherical eye model.
• To integrate collision avoidance with goal-directed behavior for realistic navigation.

Main Methods:

• Developed a model to extract depth structure from translational optic flow using local properties of a spherical eye.
• Implemented a collision avoidance algorithm using this depth information.
• Tested the algorithm with geometrically determined optic flow and bio-inspired correlation-type elementary motion detectors.
• Combined the collision avoidance algorithm with a goal-direction mechanism.

Main Results:

• Demonstrated that inferred depth information from optic flow is sufficient for collision avoidance.
• The algorithm successfully achieved collision avoidance using bio-inspired motion detectors.
• The model replicated key characteristics of insect collision avoidance behavior.
• Simulated agents exhibited goal-directed navigation in cluttered environments.

Conclusions:

• The proposed model provides an effective mechanism for collision avoidance in mobile agents.
• The approach successfully integrates optic flow processing with navigation strategies, mimicking insect capabilities.
• This research offers insights into biological navigation and potential applications in robotics.