Pit viper inspired robot

Academic

November 2024

Boulder

Mechanical design lead, algorithm developer

Project channel

Inspired by the dual-sensing capabilities of pit vipers—known for striking prey accurately even in complete darkness—we designed a robotic platform integrating both infrared (IR) and visual sensing. Drawing from biology, the robot mimics how pit vipers use IR for thermal tracking and vision for navigation. We built a custom differential-drive platform with a snake-head-like housing for accurate sensor placement. The robot used a Raspberry Pi Camera Module 2 for visual input and an Adafruit MLX90640 thermal sensor for IR imaging. We developed separate algorithms to extract centroids from both sensing streams and used a bang-bang control strategy to steer toward targets. IR sensing achieved the shortest time-to-reach (TTR) and highest stability in tracking, while vision-only showed the highest variance and slowest performance. Surprisingly, dual sensing did not outperform IR alone in TTR, though it did reduce heading angle variance, suggesting better directional stability. Mechanical design challenges, like poor print fit and assembly failures, were addressed using Velcro and modular design principles. The project demonstrated that IR sensing is crucial for fast, accurate target detection, while integration methods still need refinement for optimal multi-sensor fusion.