Avoidance of non-localizable obstacles in echolocating bats: A robotic model

Area of Science:

• Robotics
• Bioacoustics
• Animal Behavior

Background:

• Echolocating bats navigate complex environments with overlapping echoes, challenging their perception.
• Limited resolution of bat bio-sonar hinders obstacle localization in cluttered habitats.
• Despite localization challenges, bats demonstrate remarkable obstacle avoidance capabilities.

Purpose of the Study:

• To develop and evaluate a robotic model of bat obstacle avoidance using minimal sensory cues.
• To investigate the effectiveness of two strategies: Fixed Head and Delayed Linear Adaptive Law (DLAL) with acoustic gaze scanning.
• To assess the adaptive value of acoustic gaze scanning for obstacle avoidance when echoes are non-localizable.

Main Methods:

• A robotic model was implemented using interaural level differences and distance to the nearest obstacle.
• Two obstacle avoidance strategies were evaluated: Fixed Head and DLAL with acoustic gaze scanning.
• Robot performance was tested in cluttered environments simulating behavioral experimental setups.

Main Results:

• The robotic model successfully avoided obstacles using only interaural level differences and distance to the nearest object.
• Acoustic gaze scanning (DLAL strategy) resulted in reduced obstacle avoidance performance.
• The findings suggest gaze scanning may not be beneficial when angular information is limited.

Conclusions:

• Interaural level differences and distance to the nearest obstacle are sufficient cues for bat-like obstacle avoidance in cluttered environments.
• Acoustic gaze scanning may not be advantageous for obstacle avoidance when echoes are non-localizable.
• The study provides insights into the sensory strategies bats employ for navigation and survival.