Neural Network-Based Autonomous Search Model with Undulatory Locomotion Inspired by Caenorhabditis Elegans

Area of Science:

• Biomimetic robotics
• Computational neuroscience
• Locomotion control

Background:

• Caenorhabditis elegans (C. elegans) displays sophisticated chemotaxis using klinokinesis and klinotaxis.
• Its simple nervous system and unique locomotion inspire robotic navigation.
• Existing robotic schemes often lack the efficiency of biological chemotaxis.

Purpose of the Study:

• To develop an autonomous search model inspired by C. elegans chemotaxis.
• To integrate klinokinesis, klinotaxis, and undulatory locomotion for gradient searching.
• To create a cost-effective, single-sensor navigation system.

Main Methods:

• Evolved a central pattern generator for rhythmic signal propagation.
• Designed a minimal network unit with proprioceptive feedback for undulatory locomotion.
• Incorporated adaptive sensory neuron models with state-dependent gating for dual strategy execution.

Main Results:

• The model successfully integrated C. elegans chemotaxis strategies with undulatory locomotion.
• Simulations confirmed the model's effectiveness, superiority, and realism.
• The system efficiently searched for concentration peaks using a single sensor.

Conclusions:

• The developed model effectively mimics C. elegans chemotaxis for autonomous navigation.
• This approach offers a theoretical prototype for worm-like robots navigating environmental gradients.
• The model demonstrates a simple yet powerful method for cost-effective robotic search strategies.