Overview

The Objective

The objective was to design and prototype a mobile system capable of observing wildlife without requiring a nearby human operator. The robot needed to remotely navigate uneven terrain and confined environments while providing a live camera feed and identifying wildlife through visual and audio sensing.

The system also needed to remain compact, quiet, power-efficient and visually unobtrusive to minimise its effect on both wildlife and the surrounding environment. Operators controlled and monitored the robot from a separate room with no direct line of sight to the test arena.

Development

System Architecture and Milestones

Development of the system was divided into several interconnected areas:

• Mechanical Design & Rapid Prototyping: RatBot was designed as a modular platform comprising a chassis, custom tracked drive and extending camera arm. Almost all structural components were manufactured through 3D printing, using PLA for general components and ABS where greater strength was required. This allowed each subsystem to be rapidly redesigned, tested and integrated throughout development.

  
  

• Tracked Drive System: A differential track system was developed to provide traction, stability and zero-radius turning across slopes, bumps and confined passages. Each track used individually printed links, rubber contact pads, steel connecting rods and spring-loaded idlers to maintain ground contact over uneven terrain.

  
  

• Extending Camera Arm: A universal-joint camera arm provided three degrees of freedom: pan, tilt and linear extension. Two servos controlled the joint through wire linkages, while a rack-and-pinion mechanism extended the camera to inspect areas that the main robot body could not reach.

  
  

• Remote Control & Motor Feedback: A Raspberry Pi 4 acted as the robot’s onboard controller. Python clients communicated with the robot through TCP sockets, transmitting movement, motor-speed, camera-arm and lighting commands every 50 milliseconds. Encoder feedback and calibrated PD controllers were used to regulate the two drive motors.

  
  

• GUI & Sensor Integration: The robotics system was integrated with an electrical-team payload capable of real-time visual and audio wildlife recognition. The combined GUI displayed the live camera feed, detection outputs, connection controls, operator instructions and live robot-control information.

  
  

• Testing & Verification: The completed robot passed 17 of its 19 verification tests. It successfully traversed a 30-degree incline with raised obstacles, reached a maximum speed of 27 cm/s and achieved an average communication latency of 14 ms. The final platform weighed 835.7 g and cost $260.74 to construct.

The final system successfully completed all three stages of the arena and demonstrated the viability of a compact, low-cost platform for remote ecological monitoring. The main limitations identified were network congestion, reduced live-stream frame rate and inconsistent motor control at very low speeds.