22 April 2024 in IEEE Robotics and Automation Letters
Topological maps are favorable for their small storage compared to geometric maps. However, they are limited in relocalization and path planning capabilities. To solve the problem, a feature-based hybrid topological map (FHT-Map) is proposed along with a real-time map construction algorithm based on robot exploration. Specifically, the FHT-Map utilizes both RGB cameras and LiDAR information and consists of two types of nodes: main node and support node. Main nodes store visual information compressed by convolutional neural network and local laser scan data to enhance subsequent relocalization capability. Support nodes retain a minimal amount of data to ensure storage efficiency while facilitating path planning. After map construction through robot exploration, the FHT-Map can be used by other robots for relocalization and path planning. Simulation results demonstrate that the proposed FHT-Map can effectively improve relocalization and path planning capability compared with other topological maps. Moreover, simulations of the hybrid architecture are implemented to show the necessity of two types of nodes.
6 February 2025 in IEEE Transactions on Robotics
The containment of multi-robot systems (MRSs) has a wide range of applications. However, time delays in communication among robots introduce difficulties to the system to accomplish containment. Additionally, the specific dynamics of robot models poses new nonlinear and nonholonomic challenges. To solve these problems, a containment control law is proposed first for double- integrator MRSs subject to non-uniform time-varying delays. In contrast to impractical uniform delays, non-uniform time-varying delays are considered more deeply from the perspective of the Laplace matrix in this paper. The stability is proved by the Lyapunov-Krasovskii function and linear matrix inequalities. The proposed control law is further refined into a dual-loop structure for multi-nonholonomic-mobile-robot systems (NHMRSs), addressing the problem of nonholonomic constraints. Specifically, the first loop decouples the control inputs in a finite time, and then the nonholonomic robot models are regarded as linear models, which facilitates the proof of system stability. The effectiveness of the above two control laws is validated through simulations and experiments. Under these containment control laws, followers in the system reach the convex hull formed by leaders and meet the convergence objective despite the constraint of non-uniform time-varying delays.
3 April 2025 in IEEE Robotics and Automation Letters
During complex object manipulation, manipulator systems often face the configuration disconnectivity problem due to closed-chain constraints. Although regrasping can be adopted to guarantee connectivity, it introduces additional issues such as impact and efficiency. Therefore, regrasping numbers should be minimized during manipulation. To address this problem, a novel planning framework is proposed for multiple mobile manipulator systems. Given the object trajectory and the grasping pose set, the planning framework includes three steps. First, the inverse kinematic solution is verified along the given trajectory based on different grasping poses. Coverable trajectory segments are determined for each robot for a specific grasping pose. Second, the trajectory choice problem is formulated into a set cover problem, by which we can quickly determine whether the manipulation can be completed without regrasping or with a minimal regrasping number. Finally, the motions of each mobile manipulator are planned with the assigned trajectory segments using existing methods. Both simulations and experimental results show the performance of the planner in complex flipping manipulation. Additionally, theoretical analysis and multiple simulations are conducted to demonstrate the performance of the planner.
13 September 2024 in IEEE Robotics and Automation Letters
Rendezvous aims at gathering all robots at a specific location, which is an important collaborative behavior for multi-robot systems. However, in an unknown environment, it is challenging to achieve rendezvous. Previous researches mainly focus on special scenarios where communication is not allowed and each robot executes a random searching strategy, which is highly time-consuming, especially in large-scale environments. In this letter, we focus on rendezvous in unknown environments where communication is available. We divide this task into two steps: rendezvous based environment exploration with relative pose (RP) estimation and rendezvous point selection. A new strategy called partitioned and incomplete exploration for rendezvous (PIER) is proposed to efficiently explore the unknown environment, where lightweight topological maps are constructed and shared among robots for RP estimation with very few communications. Then, a rendezvous point selection algorithm based on the merged topological map is proposed for efficient rendezvous for multi-robot systems. The effectiveness of the proposed methods is validated in both simulations and real-world experiments.
28 May 2024 in IEEE Robotics and Automation Letters
Mutual localization serves as the foundation for collaborative perception in multi-robot systems. Effectively utilizing limited onboard sensors for mutual localization between marker-less robots is worthwhile. However, due to inadequate consideration of large scale variations of the robot and localization refinement, previous work has shown limited accuracy when robots are equipped only with RGB cameras. To enhance the precision of localization, this letter proposes a novel rendezvous-based hierarchical architecture for mutual localization (RHAML). Firstly, to learn multi-scale features, anisotropic convolutions are introduced into the network, yielding initial localization results. Then, the iterative refinement module is employed to adjust the poses. Finally, the pose graph is conducted to globally optimize localization results, which takes into account multi-frame observations. Therefore, a flexible architecture is provided that allows for the selection of appropriate modules based on requirements. Simulations demonstrate that RHAML effectively addresses the problem of multi-robot mutual localization, achieving translation errors below 2 cm and rotation errors below 0.5 degrees when robots exhibit 5 m of depth variation. Moreover, its practical utility is validated by applying it to map fusion when multi-robots explore unknown environments.
16 January 2026 in IEEE Transactions on Instrumentation and Measurement
Relative Localization (RelLoc) is crucial for consistent decision-making in multi-robot systems. To achieve efficient RelLoc, two key problems must be solved: how to evaluate the RelLoc accuracy, and how to coordinate the robot swarm to enhance it when it does not meets the stopping criterion. Existing research mainly focuses on passively optimizing relative poses, without explicitly evaluating the uncertainty in this process or efficiently improving RelLoc accuracy in an active manner. In this paper, we propose a novel Dynamic Localization Graph (DLG) to evaluate, monitor, and actively improve inter-robot RelLoc accuracy. To address the first problem, we propose the Metric of Relative Localization Accuracy (MRLA) based on the pose covariance without requiring prior knowledge of relative poses. Specifically, graduated non-convexity factor graph optimization and uncertainty propagation is applied to calculate the transformation and update the pose covariance between the initial frames, followed by outlier rejection through hypothesis testing. Then, the DLG is constructed by assigning the MRLA between robots as the edge weight, which illustrates the status of RelLoc in the multi-robot system. To address the second problem, we further introduce an active RelLoc enhancement strategy on top of the DLG that triggers additional inter-robot measurements when needed, enabling autonomous, accuracy-cost aware RelLoc enhancement. When the MRLA between all robot pairs exceeds the threshold, the stopping criterion for RelLoc is satisfied, ensuring high-accuracy RelLoc among robots. The DLG is sensor-independent, focusing on evaluating and enhancing RelLoc accuracy. It is instantiated on a place recognition RelLoc pipeline as a representative case study. Simulations demonstrate that when the MRLA in the DLG reaches 0.98, the Absolute Trajectory Error is less than 0.1 m, and the DLG is further verified in both indoor and outdoor experiments.