Reinforcement Learning with Elastic Time Steps

Traditional Reinforcement Learning (RL) algorithms are usually applied in robotics to learn controllers that act with a fixed control rate. Given the discrete nature of RL algorithms, they are oblivious to the effects of the choice of control rate: finding the correct control rate can be difficult and mistakes often result in excessive use of computing resources or even lack of convergence. We propose Soft Elastic Actor-Critic (SEAC), a novel off-policy actor-critic algorithm to address this issue. SEAC implements elastic time steps, time steps with a known, variable duration, which allow the agent to change its control frequency to adapt to the situation. In practice, SEAC applies control only when necessary, minimizing computational resources and data usage. We evaluate SEAC's capabilities in simulation in a Newtonian kinematics maze navigation task and on a 3D racing video game, Trackmania. SEAC outperforms the SAC baseline in terms of energy efficiency and overall time management, and most importantly without the need to identify a control frequency for the learned controller. SEAC demonstrated faster and more stable training speeds than SAC, especially at control rates where SAC struggled to converge. We also compared SEAC with a similar approach, the Continuous-Time Continuous-Options (CTCO) model, and SEAC resulted in better task performance. These findings highlight the potential of SEAC for practical, real-world RL applications in robotics.

Deployable Reinforcement Learning with Variable Control Rate

Deploying controllers trained with Reinforcement Learning (RL) on real robots can be challenging: RL relies on agents' policies being modeled as Markov Decision Processes (MDPs), which assume an inherently discrete passage of time. The use of MDPs results in that nearly all RL-based control systems employ a fixed-rate control strategy with a period (or time step) typically chosen based on the developer's experience or specific characteristics of the application environment. Unfortunately, the system should be controlled at the highest, worst-case frequency to ensure stability, which can demand significant computational and energy resources and hinder the deployability of the controller on onboard hardware. Adhering to the principles of reactive programming, we surmise that applying control actions only when necessary enables the use of simpler hardware and helps reduce energy consumption. We challenge the fixed frequency assumption by proposing a variant of RL with variable control rate. In this approach, the policy decides the action the agent should take as well as the duration of the time step associated with that action. In our new setting, we expand Soft Actor-Critic (SAC) to compute the optimal policy with a variable control rate, introducing the Soft Elastic Actor-Critic (SEAC) algorithm. We show the efficacy of SEAC through a proof-of-concept simulation driving an agent with Newtonian kinematics. Our experiments show higher average returns, shorter task completion times, and reduced computational resources when compared to fixed rate policies.

PEACE: Prompt Engineering Automation for CLIPSeg Enhancement in Aerial Robotics

From industrial to space robotics, safe landing is an essential component for flight operations. With the growing interest in artificial intelligence, we direct our attention to learning based safe landing approaches. This paper extends our previous work, DOVESEI, which focused on a reactive UAV system by harnessing the capabilities of open vocabulary image segmentation. Prompt-based safe landing zone segmentation using an open vocabulary based model is no more just an idea, but proven to be feasible by the work of DOVESEI. However, a heuristic selection of words for prompt is not a reliable solution since it cannot take the changing environment into consideration and detrimental consequences can occur if the observed environment is not well represented by the given prompt. Therefore, we introduce PEACE (Prompt Engineering Automation for CLIPSeg Enhancement), powering DOVESEI to automate the prompt generation and engineering to adapt to data distribution shifts. Our system is capable of performing safe landing operations with collision avoidance at altitudes as low as 20 meters using only monocular cameras and image segmentation. We take advantage of DOVESEI's dynamic focus to circumvent abrupt fluctuations in the terrain segmentation between frames in a video stream. PEACE shows promising improvements in prompt generation and engineering for aerial images compared to the standard prompt used for CLIP and CLIPSeg. Combining DOVESEI and PEACE, our system was able improve successful safe landing zone selections by 58.62% compared to using only DOVESEI. All the source code is open source and available online.

Dynamic Open Vocabulary Enhanced Safe-landing with Intelligence (DOVESEI)

This work targets what we consider to be the foundational step for urban airborne robots, a safe landing. Our attention is directed toward what we deem the most crucial aspect of the safe landing perception stack: segmentation. We present a streamlined reactive UAV system that employs visual servoing by harnessing the capabilities of open vocabulary image segmentation. This approach can adapt to various scenarios with minimal adjustments, bypassing the necessity for extensive data accumulation for refining internal models, thanks to its open vocabulary methodology. Given the limitations imposed by local authorities, our primary focus centers on operations originating from altitudes of 100 meters. This choice is deliberate, as numerous preceding works have dealt with altitudes up to 30 meters, aligning with the capabilities of small stereo cameras. Consequently, we leave the remaining 20m to be navigated using conventional 3D path planning methods. Utilizing monocular cameras and image segmentation, our findings demonstrate the system's capability to successfully execute landing maneuvers at altitudes as low as 20 meters. However, this approach is vulnerable to intermittent and occasionally abrupt fluctuations in the segmentation between frames in a video stream. To address this challenge, we enhance the image segmentation output by introducing what we call a dynamic focus: a masking mechanism that self adjusts according to the current landing stage. This dynamic focus guides the control system to avoid regions beyond the drone's safety radius projected onto the ground, thus mitigating the problems with fluctuations. Through the implementation of this supplementary layer, our experiments have reached improvements in the landing success rate of almost tenfold when compared to global segmentation. All the source code is open source and available online (github.com/MISTLab/DOVESEI).

Congestion and Scalability in Robot Swarms: a Study on Collective Decision Making

One of the most important promises of decentralized systems is scalability, which is often assumed to be present in robot swarm systems without being contested. Simple limitations, such as movement congestion and communication conflicts, can drastically affect scalability. In this work, we study the effects of congestion in a binary collective decision-making task. We evaluate the impact of two types of congestion (communication and movement) when using three different techniques for the task: Honey Bee inspired, Stigmergy based, and Division of Labor. We deploy up to 150 robots in a physics-based simulator performing a sampling mission in an arena with variable levels of robot density, applying the three techniques. Our results suggest that applying Division of Labor coupled with versioned local communication helps to scale the system by minimizing congestion.

BittyBuzz: A Swarm Robotics Runtime for Tiny Systems

Swarm robotics is an emerging field of research which is increasingly attracting attention thanks to the advances in robotics and its potential applications. However, despite the enthusiasm surrounding this area of research, software development for swarm robotics is still a tedious task. That fact is partly due to the lack of dedicated solutions, in particular for low-cost systems to be produced in large numbers and that can have important resource constraints. To address this issue, we introduce BittyBuzz, a novel runtime platform: it allows Buzz, a domain-specific language, to run on microcontrollers while maintaining dynamic memory management. BittyBuzz is designed to fit a flash memory as small as 32 kB (with usable space for scripts) and work with as little as 2 kB of RAM. In this work, we introduce the BittyBuzz implementation, its differences from the original Buzz virtual machine, and its advantages for swarm robotics systems. We show that BittyBuzz is successfully integrated with three robotic platforms with minimal memory footprint and conduct experiments to show computation performance of BittyBuzz. Results show that BittyBuzz can be effectively used to implement common swarm behaviors on microcontroller-based systems.

Joint Machine-Transporter Scheduling for Multistage Jobs with Adjustable Computation Time

This paper presents a scalable solution with adjustable computation time for the joint problem of scheduling and assigning machines and transporters for missions that must be completed in a fixed order of operations across multiple stages. A battery-operated multi-robot system with a maximum travel range is employed as the transporter between stages and charging them is considered as an operation. Robots are assigned to a single job until its completion. Additionally, The operation completion time is assumed to be dependent on the machine and the type of operation, but independent of the job. This work aims to minimize a weighted multi-objective goal that includes both the required time and energy consumed by the transporters. This problem is a variation of the flexible flow shop with transports, that is proven to be NP-complete. To provide a solution, time is discretized, the solution space is divided temporally, and jobs are clustered into diverse groups. Finally, an integer linear programming solver is applied within a sliding time window to determine assignments and create a schedule that minimizes the objective. The computation time can be reduced depending on the number of jobs selected at each segment, with a trade-off on optimality. The proposed algorithm finds its application in a water sampling project, where water sampling jobs are assigned to robots, sample deliveries at laboratories are scheduled, and the robots are routed to charging stations.

Energy Sufficiency in Unknown Environments via Control Barrier Functions

Maintaining energy sufficiency of a battery-powered robot system is a essential for long-term missions. This capability should be flexible enough to deal with different types of environment and a wide range of missions, while constantly guaranteeing that the robot does not run out of energy. In this work we present a framework based on Control Barrier Functions (CBFs) that provides an energy sufficiency layer that can be applied on top of any path planner and provides guarantees on the robot's energy consumption during mission execution. In practice, we smooth the output of a generic path planner using double sigmoid functions and then use CBFs to ensure energy sufficiency along the smoothed path, for robots described by single integrator and unicycle kinematics. We present results using a physics-based robot simulator, as well as with real robots with a full localization and mapping stack to show the validity of our approach.

CoNi-MPC: Cooperative Non-inertial Frame Based Model Predictive Control

This paper presents a novel solution for UAV control in cooperative multi-robot systems, which can be used in various scenarios such as leader-following, landing on a moving base, or specific relative motion with a target. Unlike classical methods that tackle UAV control in the world frame, we directly control the UAV in the target coordinate frame, without making motion assumptions about the target. In detail, we formulate a non-linear model predictive controller of a UAV within a non-inertial frame (i.e., the target frame). The system requires the relative states (pose and velocity), the angular velocity and the accelerations of the target, which can be obtained by relative localization methods and ubiquitous MEMS IMU sensors, respectively. This framework eliminates dependencies that are vital in classical solutions, such as accurate state estimation for both the agent and target, prior knowledge of the target motion model, and continuous trajectory re-planning for some complex tasks. We have performed extensive simulations to investigate the control performance considering the varying motion characteristics of the target. Furthermore, we conducted considerable real robot experiments, employing laboratory motion-capture systems or relative localization methods implemented outdoors, to validate the applicability and feasibility of the proposed approach.

Real-Time Simultaneous Localization and Mapping with LiDAR intensity

We propose a novel real-time LiDAR intensity image-based simultaneous localization and mapping method , which addresses the geometry degeneracy problem in unstructured environments. Traditional LiDAR-based front-end odometry mostly relies on geometric features such as points, lines and planes. A lack of these features in the environment can lead to the failure of the entire odometry system. To avoid this problem, we extract feature points from the LiDAR-generated point cloud that match features identified in LiDAR intensity images. We then use the extracted feature points to perform scan registration and estimate the robot ego-movement. For the back-end, we jointly optimize the distance between the corresponding feature points, and the point to plane distance for planes identified in the map. In addition, we use the features extracted from intensity images to detect loop closure candidates from previous scans and perform pose graph optimization. Our experiments show that our method can run in real time with high accuracy and works well with illumination changes, low-texture, and unstructured environments.