Current lidar and camera-based solutions for low-cost indoor mobile robots have limitations such as poor performance in visually obscured environments, high computational overhead for data processing, and high costs for lidars. In contrast, mmWave radar sensors offer a cost-effective and lightweight alternative, providing accurate ranging regardless of visibility. However, existing radar-based localization suffers from sparse point cloud generation, noise, and false detections. Thus, in this work, we introduce RaGNNarok, a real-time, lightweight, and generalizable graph neural network (GNN)-based framework to enhance radar point clouds, even in complex and dynamic environments. With an inference time of only 7.3 ms on the low-cost Raspberry Pi 5, RaGNNarok runs even on such resource-constrained devices, without additional computational resources. We evaluate its performance across key tasks, including localization, SLAM, and autonomous navigation, in three different environments. Our results demonstrate strong reliability and generalizability, making RaGNNarok a robust solution for low-cost indoor mobile robots.

In this work, we introduce REFORMA, a novel robust reinforcement learning (RL) approach to design controllers for unmanned aerial vehicles (UAVs) robust to unknown disturbances during flights. These disturbances, typically due to wind turbulence, electromagnetic interference, temperature extremes and many other external physical interference, are highly dynamic and difficult to model. REFORMA can perform a real-time online adaptation to these disturbances and generate appropriate velocity actions as countermeasures to stabilize the drone. REFORMA consists of two components: a base policy trained completely in simulation using model-free RL and an adaptation module trained via supervised learning with on-policy datasets. By varying the disturbance strength in an adaptation module, i.e., adopting adaptive adversary, the policy is then able to handle extreme cases when the velocity of the drone is immediately affected by disturbances. Finally, we demonstrate the effectiveness of our method through extensive simulated experiments. To the best of our knowledge, REFORMA is the first robust RL approach that uses adaptive adversaries to tackle uncertain disturbances in drone tasks.

In this work, we present RadCloud, a novel real-time framework for directly obtaining higher-resolution lidar-like 2D point clouds from low-resolution radar frames on resource-constrained platforms commonly used in unmanned aerial and ground vehicles (UAVs and UGVs, respectively); such point clouds can then be used for accurate environmental mapping, navigating unknown environments, and other robotics tasks. While high-resolution sensing using radar data has been previously reported, existing methods cannot be used on most UAVs, which have limited computational power and energy; thus, existing demonstrations focus on offline radar processing. RadCloud overcomes these challenges by using a radar configuration with 1/4th of the range resolution and employing a deep learning model with 2.25× fewer parameters. Additionally, RadCloud utilizes a novel chirp-based approach that makes obtained point clouds resilient to rapid movements (e.g., aggressive turns or spins) that commonly occur during UAV flights. In real-world experiments, we demonstrate the accuracy and applicability of RadCloud on commercially available UAVs and UGVs, with off-the-shelf radar platforms on-board.

We demonstrate RadCloud, a real-time framework for obtaining high-resolution lidar-like 2D point clouds from low-resolution millimeter-wave (mmWave) radar data on resource-constrained platforms commonly found on unmanned aerial and ground vehicles (UAVs and UGVs). Such point clouds can then be used for mapping key features of the environment, route planning and navigation, and other robotics tasks. Rad-Cloud is specifically optimized for UAVs and UGVs by using a radar configuration with 1/4th the range resolution, using a model with 2.25× fewer parameters, and reducing total sensing time by a factor of 250×. The real-time ROS framework will be demonstrated on a UGV and UAV equipped with CPU-only compute platforms in diverse environments.

Harmful marine spills, such as algae blooms and oil spills, damage ecosystems and threaten public health tremendously. Hence, an effective spill coverage and removal strategy will play a significant role in environmental protection. In recent years, low-cost water surface robots have emerged as a solution, with their efficacy verified at small scale. However, practical limitations, such as connectivity, scalability, and sensing and operation ranges significantly impair their large-scale use. To circumvent these limitations, we propose a novel asymptotic boundary shrink control strategy that enables collective coverage of a spill by autonomous robots featuring customized operation ranges. For each robot, a novel controller is implemented that relies only on local vision sensors with limited vision range. Moreover, the distributedness of this strategy allows any number of robots to be employed without inter-robot collisions. Finally, features of this approach including the convergence of robot motion during boundary shrink control, spill clearance rate, and the capability to work under limited ranges of vision and wireless connectivity are validated through extensive experiments with simulation.

Multirobot rendezvous control and coordination strategies have garnered significant interest in recent years because of their potential applications in decentralized tasks. In this paper, we introduce a coordinate-free rendezvous control strategy to enable multiple robots to gather at different locations (dynamic leader robots) by tracking their hierarchy in a connected interaction graph. A key novelty in this strategy is the gathering of robots in different groups rather than at a single consensus point, motivated by autonomous multipoint recharging and flocking control problems. We show that the proposed rendezvous strategy guarantees convergence and maintains connectivity while accounting for practical considerations such as robots with limited speeds and an obstacle-rich environment. The algorithm is distributed and handles minor faults such as a broken immobile robot and a sudden link failure. In addition, we propose an approach that determines the locations of rendezvous points based on the connected interaction topology and indirectly optimizes the total energy consumption for rendezvous in all robots. Through extensive experiments with the Robotarium multirobot testbed, we verified and demonstrated the effectiveness of our approach and its properties.

Water pollution has caused increased incidence of algal growth around the globe. Harmful algae blooms result in massive economic losses. In this paper, a multi-robot based task planner is designed to remove excessive algae from water bodies and to identify algae build-up so that prompt action can be taken against its accumulation. Computer vision is incorporated to enable algae detection and area estimation based on training, comparing, and evaluating various advanced deep learning models using our custom algae dataset. We further propose a novel algorithm for robot resource allocation between bounding boxes of detected algae based on multivariable optimization. This systematic solution is evaluated in a simulated environment, demonstrating how the robots are optimally assigned to the detected algae patches for algae removal.

This paper presents a configurable image sensor with energy harvesting capabilities. The image sensor's pixels can be configured as photo-sensors for image acquisition or as tiny solar cells for energy harvesting. The proposed configurable pixel design requires four pixels and uses a N-well/P-sub photodiode which is typically regarded as a parasitic diode for energy harvesting. A chip containing a 64×64 array of configurable pixels, read-out circuitry and biasing was designed and fabricated on a standard 0.5 μm CMOS process. The pixel size is 36 μm×40.8 μm with a fill factor of 42%. Test results are presented demonstrating the dual functionality of the image sensor. Measurements show that 55 nW can be harvested for an illuminance of 500 lux at the lens plane.