RB5 Low-Cost Explorer: Implementing Autonomous Long-Term Exploration on Low-Cost Robotic Hardware

Adam Seewald, Marvin Chancán, Connor M. McCann, Seonghoon Noh, Omeed Fallahi, Hector Castillo, Ian Abraham, Aaron M. Dollar

Abstract This systems paper presents the implementation and design of RB5, a wheeled robot for autonomous long-term exploration with fewer and cheaper sensors. Requiring just an RGB-D camera and low-power computing hardware, the system consists of an experimental platform with rocker-bogie suspension. It operates in unknown and GPS-denied environments and on indoor and outdoor terrains. The exploration consists of a methodology that extends frontier- and sampling-based exploration with a path-following vector field and a state-of-the-art SLAM algorithm. The methodology allows the robot to explore its surroundings at lower update frequencies, enabling the use of lower-performing and lower-cost hardware while still retaining good autonomous performance. The approach further consists of a methodology to interact with a remotely located human operator based on an inexpensive long-range and low-power communication technology from the internet-of-things domain (i.e., LoRa) and a customized communication protocol. The results and the feasibility analysis show the possible applications and limitations of the approach.

Introduction

Autonomous robotic exploration is typically an expensive endeavor. State-of-the-art systems often rely on costly 3D LiDARs, high-performance GPUs, and complex mechanical designs, such as quadrupeds, to navigate unknown, GPS-denied environments. These requirements create a significant barrier for applications in environmental conservation, education, and large-scale surveying.

The RB5 Low-Cost Explorer addresses this challenge by "doing more with less," focusing on reducing computational intensity and energy consumption. The goal is a system capable of long-duration exploration across diverse terrains without the need for expensive sensors or high-frequency control loops that rapidly drain batteries.

System Design and Hardware Innovation

The RB5 platform is built upon a rocker-bogie suspension system, a design famously utilized by NASA’s Mars rovers. This mechanical choice is pivotal; it allows the robot to passively adapt to uneven terrain and clear obstacles up to 19 cm high. By handling terrain challenges mechanically, the system avoids the need for power-hungry active suspension or complex gait planning algorithms, significantly extending operational life.

The RB5 experimental platform showcasing the rocker-bogie suspension and the coordinate frames used for navigation.

For sensing, the RB5 eschews expensive LiDAR in favor of a single Intel RealSense D435 RGB-D camera. Localization and mapping are handled by RTAB-Map, a robust SLAM algorithm capable of processing visual and depth data to build 3D maps of unknown environments. The onboard computing is powered by an NVIDIA Jetson Xavier NX, a low-power embedded module that balances performance with energy efficiency.

Autonomous Navigation via Path-Following Vector Fields

The core software innovation of the RB5 is its hybrid exploration methodology. It combines frontier-based exploration—which identifies the boundaries between known and unknown space—with sampling-based path planning. To bridge the gap between high-level planning and low-level control on constrained hardware, the authors implemented a path-following vector field approach.

Illustration of the path-following vector field methodology at different time steps, showing how the robot navigates around obstacles.

Borrowed from aerial robotics, this technique guides the robot along a smooth manifold using gradient descent rather than constantly recalculating discrete waypoints. This approach is inherently stable even at low update frequencies (1–10 Hz). This stability is crucial for low-cost hardware, as it allows the robot to maintain autonomous performance even when the CPU/GPU is heavily loaded with SLAM or communication tasks.

Communication and Remote Intervention

Acknowledging that low-cost systems may encounter "unsolvable" edge cases, the RB5 integrates LoRa (Long Range) technology for remote intervention. LoRa provides a low-power, long-distance communication link (up to 5km), bypassing the range limitations of standard Wi-Fi. This serves as a robust safety net, allowing a human operator to intervene if the robot becomes stuck in complex environments without requiring a pre-existing mesh network infrastructure.

Experimental Results

The system was evaluated through real-world experiments in structured indoor hallways and unstructured underground tunnels. Key findings include:

  • Exceptional Battery Life: The RB5 achieved a median battery life of approximately 8.3 hours during typical operation, extending up to 16.5 hours when stationary. This significantly outperforms many high-end legged or aerial platforms.
  • Robust Navigation: The robot successfully mapped and navigated circuits over 100 meters long, demonstrating effective obstacle avoidance and path following in GPS-denied areas.
  • Computational Efficiency: The vector field approach proved effective at update rates as low as 1 Hz, validating the design goal of using lower-performing hardware without sacrificing autonomy.

Conclusion and Future Work

The RB5 demonstrates that mechanical ingenuity and efficient algorithmic design can compensate for lower-cost hardware. While the use of RGB-D cameras results in lower-density point clouds compared to LiDAR, the system's overall performance and longevity make it a viable candidate for long-term exploration tasks. Future work will focus on energy-aware path planning and enhancing the custom LoRa protocol with feature tracking to further improve remote interaction capabilities.

Summary updated

Abstract

This systems paper presents the implementation and design of RB5, a wheeled robot for autonomous long-term exploration with fewer and cheaper sensors. Requiring just an RGB-D camera and low-power computing hardware, the system consists of an experimental platform with rocker-bogie suspension. It operates in unknown and GPS-denied environments and on indoor and outdoor terrains. The exploration consists of a methodology that extends frontier- and sampling-based exploration with a path-following vector field and a state-of-the-art SLAM algorithm. The methodology allows the robot to explore its surroundings at lower update frequencies, enabling the use of lower-performing and lower-cost hardware while still retaining good autonomous performance. The approach further consists of a methodology to interact with a remotely located human operator based on an inexpensive long-range and low-power communication technology from the internet-of-things domain (i.e., LoRa) and a customized communication protocol. The results and the feasibility analysis show the possible applications and limitations of the approach.
BibTeX
@inproceedings{seewald2024rb,
  author = {Adam Seewald and Marvin Chancán and Connor M. McCann and Seonghoon Noh and Omeed Fallahi and Hector Castillo and Ian Abraham and Aaron M. Dollar},
  title = {{RB5 Low-Cost Explorer: Implementing Autonomous Long-Term Exploration on Low-Cost Robotic Hardware}},
  booktitle = {IEEE International Conference on Robotics and Automation},
  year = {2024},
  url = {https://openreview.net/forum?id=Oy8JQuJPIs},
}
Figure