BinWalker: Development and Field Evaluation of a Quadruped Manipulator Platform for Sustainable Litter Collection
Giulio Turrisi *, Angelo Bratta *, Giovanni Minelli, Gabriel Fischer Abati, Amir H. Rad, João Carlos Virgolino Soares, Claudio Semini · 2026
Read BinWalker: Development and Field Evaluation of a Quadruped Manipulator Platform for Sustainable Litter Collection by Giulio Turrisi *, Angelo Bratta *, Giovanni Minelli, Gabriel Fischer Abati, Amir H. Rad, João Carlos Virgolino Soares, Claudio Semini (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-10529.txt.
Introduction
Giulio Turrisi and co-authors present BinWalker, a four-legged robot with an arm, built to pick up litter in hard outdoor places. Litter gathers on rough ground, coasts, parks, and roadsides, and manual cleanup is slow, costly work that leaves much behind. The robot pairs legged mobility with a manipulator arm and an onboard bin, so it can spot, grasp, and store items while crossing uneven terrain. Field tests outdoors showed the approach works, a step toward helping large cleanup efforts where wheeled machines struggle.
Catalog record
| CREATOR | Giulio Turrisi *, Angelo Bratta *, Giovanni Minelli, Gabriel Fischer Abati, Amir H. Rad, João Carlos Virgolino Soares, Claudio Semini |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30101 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.10529
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
BinWalker: Development and Field Evaluation of a Quadruped Manipulator Platform for Sustainable Litter Collection Thanks: *Equal contribution Thanks: This work is part of the “Technologies for Sustainability” Flagship program of the Istituto Italiano di Tecnologia (IIT). Additional funding by the European Union - NextGenerationEU and by the Ministry of University and Research (MUR), National Recovery and Resilience Plan (NRRP), Mission 4, Component 2, Investment 1.5, project “RAISE - Robotics and AI for Socio-economic Empowerment” (ECS00000035). Giulio Turrisi *, Angelo Bratta *, Giovanni Minelli, Gabriel Fischer Abati, Amir H. Rad, João Carlos Virgolino Soares, Claudio Semini Abstract Litter pollution represents a growing environmental problem affecting natural and urban ecosystems worldwide. Waste discarded in public spaces often accumulates in areas that are difficult to access, such as uneven terrains, coastal environments, parks, and roadside vegetation. Over time, these materials degrade and release harmful substances, including toxic chemicals and microplastics, which can contaminate soil and water and pose serious threats to wildlife and human health. Despite increasing awareness of the problem, litter collection is still largely performed manually by human operators, making large-scale cleanup operations labor-intensive, time-consuming, and costly. Robotic solutions have the potential to support and partially automate environmental cleanup tasks. In this work, we present a quadruped robotic system designed for autonomous litter collection in challenging outdoor scenarios. The robot combines the mobility advantages of legged locomotion with a manipulation system consisting of a robotic arm and an onboard litter container. This configuration enables the robot to detect, grasp, and store litter items while navigating through uneven terrains. The proposed system aims to demonstrate the feasibility of integrating perception, locomotion, and manipulation on a legged robotic platform for environmental cleanup tasks. Experimental evaluations conducted in outdoor scenarios highlight the effectiveness of the approach and its potential for assisting large-scale litter removal operations in environments that are difficult to reach with traditional robotic platforms. The code associated with this work can be found at: https://github.com/iit-DLSLab/trash-collection-isaaclab . Fig. 1: Prototype based on the Aliengo robot. Top: Isometric view of the robot highlighting the components mounted on the top of the trunk, including the frontal camera, the Z1 arm, and the litter container. Bottom: detailed view of the container opening mechanism used to unload the collected litter. Fig. 2: Block scheme of the proposed approach. I Introduction Litter pollution represents a growing environmental challenge affecting both urban and natural ecosystems [ 22 ] . Waste discarded in public areas often accumulates in locations such as parks, coastal regions, and roadside environments, where it may persist for long periods and progressively degrade into harmful substances, including toxic chemicals and microplastics. These pollutants can contaminate the soil and water and pose a serious threat to wildlife and human health. Despite increasing awareness of this issue, litter removal is still largely performed manually by humans [ 13 ] , making large-scale cleanup operations labor-intensive, costly, and difficult to sustain over time. Robotic systems have recently emerged as a promising solution to support environmental cleanup tasks. Several autonomous platforms have been proposed for litter detection and collection in outdoor environments [ 10 ] . Most of these systems are based on wheeled or tracked robots operating in structured or semi-structured settings such as sidewalks, beaches, or urban pavements [ 8 ] . Although such platforms can be effective in relatively flat terrains, their mobility becomes significantly limited in environments characterized by uneven ground, vegetation, obstacles, or narrow passages. Legged robots offer an attractive alternative for these scenarios due to their superior mobility and terrain adaptability [ 1 ] . In recent years, quadruped robots have demonstrated remarkable capabilities in traversing rough terrain, climbing obstacles, and navigating complex outdoor environments [ 5 , 9 , 19 ] . These characteristics make them particularly suitable for environmental monitoring and intervention tasks in areas that are difficult to access using traditional mobile robots. Alongside mobility, perception, and manipulation capabilities are essential for autonomous litter collection. Recent advances in deep learning have significantly improved the performance of vision-based object detection systems, enabling reliable identification of litter in cluttered outdoor environments [ 18 , 17 , 7 ] . Convolutional neural networks have been successfully applied to detect various categories of litter, including plastic debris, bottles, and cigarette butts, providing a robust perception pipeline for robotic cleanup applications [ 15 , 10 ] . However, integrating perception with efficient planning and manipulation remains a challenging problem, especially when operating on mobile robotic platforms in dynamic outdoor environments. Regarding mobile manipulation, different works have studied the problem of loco-manipulation [ 14 ] , [ 26 ] , [ 23 ] under a non-specific setting. Here, instead, we tailor this problem in the context of litter-collection, with the goal of studying the challenges that are inherently related to this task. In this work, we present a quadruped robotic system designed for autonomous litter collection in challenging outdoor environments. The proposed platform combines the terrain adaptability of legged locomotion with a robotic arm and an onboard litter container, enabling the robot to detect, grasp, and store litter items while navigating through irregular terrains. Litter detection is performed using a convolutional neural network trained to identify litter in outdoor scenes. Collection is then decoupled by a hierarchical approach, separating arm and base pose optimization which relies on standard Inverse Kinematics (IK), from locomotion, which instead is developed by adopting a Reinforcement Learning (RL) algorithm to reach robust walking over challenging terrains. I-A Contribution The main contributions of this work can be summarized as follows: • A mobile robotic platform for autonomous litter collection that integrates quadruped locomotion, manipulation, perception, and onboard litter storage. • An experimental evaluation demonstrating the feasibility of autonomous litter collection using a legged robotic platform. Furthermore, this manuscript discusses practical insights, highlighting what works well and identifying aspects that should be improved to enhance the performance of autonomous robots in this task. Finally, we release the code associated with this work with the goal of providing the community with a baseline upon which their results can be compared or further research can be built. I-B Outline The paper is organized as follows: Sec. II gives an overview of our litter collection prototype, whereas Sec. III describes the proposed approach delving into the components of locomotion, manipulation, and vision. In Section IV , we report on the application of the proposed prototype in an outdoor experimental scenario, drawing discussion of current limitations and possible future work directions. II Litter Collection Hardware For this work, the commercial quadruped robot Unitree Aliengo 1 1 1 Unitree Aliengo: https://www.unitree.com/mobile/aliengo/ has been adapted by integrating a robotic manipulation and storage system to enable autonomous litter collection (see Fig. 1 -top). A 6-DoF lightweight manipulator, the Unitree Z1 robotic arm 2 2 2 Unitree Z1 Pro: https://www.unitree.com/z1 , is mounted in the frontal section of the robot trunk through a custom aluminum support frame. This configuration, coupled with the position of the stereo camera Realsense D435 under the base of the arm, allows the manipulator to operate within a large frontal workspace. On top of the trunk of the robot, right behind the manipulator arm, we placed a cabinet acting as a bin container. The cabinet consists of a