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

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.