Bio-Inspired Robotics & Research

Pit Viper-Inspired Integration of Infrared and Visual Sensing

Bio-Inspired Robotic Navigation and Target Detection

Duration
4 months
Team
Julianne Attai, Simon Jones, Tanmay Dharmaji Mhatre, Saravana Kumar Prakash
Client
University of Colorado Boulder - Robotics Program

Project Overview

This research project investigates how pit vipers leverage both infrared (IR) and visual sensing for hunting, and applies this biological inspiration to robotic navigation and target detection. A differential drive robot was designed with dual-sensing capabilities using an Adafruit MLX90640 IR Thermal Camera and Raspberry Pi Camera Module 2 to test the hypothesis that combining IR and vision will outperform single-sensing approaches in navigation and prey detection.

The Challenge

Pit vipers can strike prey even in darkness using pit organs for infrared detection, while also relying on visual cues for navigation and landmark recognition. The challenge was to design a robotic platform that could integrate both sensing modalities and test whether dual-sensing approaches outperform single-sensing methods in target tracking and navigation tasks.

Our Solution

Developed a snake-inspired mechanical platform with dual-sensing capabilities. Implemented thermal image processing using binary masking (>35°C) and centroid calculation via scipy.ndimage.center_of_mass. Created visual image processing using AprilTags for low-power recognition, calculating distance and heading from tag size and pixel location. Integrated both modalities through simple averaging of heading angles with a bang-bang controller for steering control.

Bio-Inspired Gallery

Pit Viper-Inspired Integration of Infrared and Visual Sensing - Image 1
Auto
1 / 6

Interactive Model

Interactive Assembly
eDrawings Viewer

Exploded assembly view showing component relationships

Key Results

0
Thermal Sensing Performance
Accuracy / efficiency score — Represents highest target tracking accuracy and lowest heading error (~15° variance).
0
Visual Sensing Performance
Accuracy / efficiency score — Lowest accuracy due to large heading variance (~20°) and longest time-to-reach (TTR).
0
Dual Sensing Performance
Accuracy / efficiency score — Combined sensing improved stability but was slightly slower than IR alone.
0
Target Detection Accuracy
Target detection success rate — System successfully detected and tracked target at ~1 m distance across trials.
0
System Integration
Integration effectiveness — Fusion method (simple averaging) worked but wasn’t optimal; could improve with weighted fusion.

Technologies & Tools

Raspberry Pi 4
Adafruit MLX90640 IR Thermal Camera
Raspberry Pi Camera Module 2
AprilTags
Python/OpenCV
3D Printing
Differential Drive Control
Snapfit Wheels

Project Timeline

1
1 month

Research & Design

Literature review on pit viper sensing, mechanical platform design, and sensor selection

Deliverables:
  • Research paper analysis
  • Mechanical design concepts
  • Sensor selection and specifications
2
1.5 months

Platform Development

3D printing and assembly of snake-head inspired robot platform with dual sensing capabilities

Deliverables:
  • 3D printed mechanical platform
  • Sensor integration and mounting
  • Power and control system setup
3
1 month

Algorithm Development

Development of thermal and visual image processing algorithms, sensor fusion, and control logic

Deliverables:
  • Thermal image processing algorithm
  • Visual tracking with AprilTags
  • Sensor fusion and control algorithms
4
0.5 months

Testing & Analysis

Comprehensive testing of single and dual sensing modalities with performance analysis

Deliverables:
  • Performance testing results
  • Statistical analysis of TTR and variance
  • Research findings and conclusions

Technical Specifications

Platform TypeDifferential drive robot with snake-head inspired design
Primary SensorAdafruit MLX90640 IR Thermal Camera (55° field of view)
Secondary SensorRaspberry Pi Camera Module 2
Processing UnitRaspberry Pi 4
Target RecognitionAprilTags for visual tracking
Thermal Threshold>35°C for target detection
Control SystemBang-bang controller with 1° heading tolerance
Target Distance~1m for testing
Trial Configuration5 trials each for visual-only, thermal-only, and dual-sensing

Key Features

Bio-inspired design mimicking pit viper sensory integration
Dual-sensing platform with IR thermal and visual cameras
Advanced thermal image processing with binary masking and centroid calculation
Visual tracking using AprilTags for low-power target recognition
Sensor fusion through heading angle averaging
Differential drive control system with bang-bang steering
3D printed snake-head inspired mechanical design
Comprehensive performance testing and statistical analysis
Research contribution to bio-inspired robotics field

Team

Julianne Attai

Simon Jones

Saravana Kumar Prakash

Interested in Similar Work?

Let's discuss how we can bring innovative engineering solutions to your next project.