Bio-Inspired Swarm Navigation on Resource-Constrained Robots for GPS-Denied Environments

Area of Science:

• Robotics
• Sensor Technology
• Artificial Intelligence

Background:

• Developing autonomous robotic systems for complex tasks requires sophisticated sensing and efficient computation.
• Existing swarm platforms often face limitations in detection accuracy, power consumption, and localization precision.

Purpose of the Study:

• To experimentally validate a novel robotic swarm platform integrating advanced pheromone detection capabilities.
• To quantify the performance of optical and chemical sensing modalities, computational efficiency, power consumption, and localization accuracy.
• To demonstrate the system's suitability for extended missions and coordinated multi-robot operations.

Main Methods:

• Optical and dual-modality pheromone detection with accuracy and robustness testing under various environmental conditions.
• Calibration and performance evaluation of the MQ-135 chemical sensor across temperature and humidity ranges.
• Quantification of convolutional neural network (CNN) inference speedup via 8-bit quantization and power consumption analysis.
• Measurement of peripheral subsystem and complete per-robot system power draw, battery life estimation.
• Localization accuracy assessment and multi-robot coordination tests, including map convergence and collision rate analysis.

Main Results:

• Optical pheromone detection achieved 88.7% accuracy, with the dual-modality channel at 86.1%, robust to interference.
• MQ-135 sensor calibration showed R² ≥ 0.999, with minimal voltage drift.
• CNN inference achieved 15 FPS with 3.2x speedup at 1.8 W power consumption.
• The complete robot system consumed 6.15 W, enabling six-hour missions.
• Localization error was 0.074 m, with 97.5% map coverage; multi-robot tests showed efficient convergence and low collision rates.

Conclusions:

• The validated robotic swarm platform demonstrates high performance in pheromone detection, computational efficiency, and localization.
• The system's low power consumption and robust capabilities enable extended autonomous missions.
• This work sets a new benchmark for physical swarm platforms operating under demanding constraints.