Thijs Van Hauwermeiren Van Hauwermeiren

Mr. Thijs Van Hauwermeiren Van Hauwermeiren

orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 June 2022 in IEEE/ASME Transactions on Mechatronics
Verified
Journal Article
Closed access
SWSander De Witte
Sander De Witte
THThijs Van Hauwermeiren Van Hauwermeiren
Thijs Van Hauwermeiren Van Hauwermeiren
TLTom Lefebvre
Tom Lefebvre
GCGuillaume Crevecoeur
Guillaume Crevecoeur
Cooperative multiagent manipulation systems allow to extend on the manipulative limitations of individual agents, increasing the complexity of the manipulation tasks the ensemble can handle. Controlling such a system requires meticulous planning of subsequent subtasks, queried to the individual agents, in order to execute the master task successfully. Real-time (re)planning is essential to ensure the task can still be achieved when subtasks execution suffers from uncertainty or when the master task changes intermittently requiring real-time reconfiguration of the plan. In this article, we develop a supervisory control architecture tailored to the cooperation of two robotic manipulators equipped with standard pick-and-place facilities in the plane. We control the planar position and orientation of an object using two underactuated manipulators so that only the position of the object can be controlled directly. The desired orientation follows from the accumulation of alternating relative angles. A time-invariant policy function is trained using deep reinforcement learning, which can determine a finite sequence of pick-and-place maneuvers to manipulate the object to a desired configuration. Two policy architectures are compared. The first uses the kinematic model to determine the final step, whilst the second policy makes this decision itself. The more information is given to the policy, the easier it trains. In return, it becomes less adaptable and loses some of its generalizability.

24 May 2024 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
THThijs Van Hauwermeiren Van Hauwermeiren
Thijs Van Hauwermeiren Van Hauwermeiren
ASAnatolii Sianov
Anatolii Sianov
ACAnnelies Coene
Annelies Coene
GCGuillaume Crevecoeur
Guillaume Crevecoeur
This paper presents a new tactile sensor that utilizes an array of barometric pressure sensors encased in a deformable rubber sphere. Designed as an end effector foot of the legged quadruped robot Unitree Go1, the proposed sensor is able to withstand repeated impacts of at least 40 N at the feet. Tactile sensing in legged robotics extends their utility specifically in the context of edge detection and exploration. The presented tactile contact framework processes the pressure data to classify the type of contact (no contact, flat or edge) and the orientation of the edges relative to robot base. To assess the performance of the sensors and their ability for tactile edge exploration we extensively test with the legged robot in varied conditions including different terrains, changing payloads, exposure to dynamic disturbances, and sloped edges. The edge detection is compared against the original scalar force sensors. Experiments demonstrate a mean absolute error of 2 $^{\circ }$ for predicted edge orientation at a detection range of 14 mm and robust operation in realistic operating conditions for medium-sized quadruped robots. With this contribution, we aim to enhance the capabilities and safety of legged robots in various applications.

19 May 2025 in Machines
Verified
Journal Article
Open access
PDF file
THThijs Van Hauwermeiren Van Hauwermeiren
Thijs Van Hauwermeiren Van Hauwermeiren
ACAnnelies Coene
Annelies Coene
GCGuillaume Crevecoeur
Guillaume Crevecoeur
Ground reaction forces (GRFs) are the primary interaction forces that enable a legged robot to maintain balance and perform locomotion. Most quadruped robot controllers estimate GRFs indirectly using joint torques and a kinematic model, which depend on assumptions and are highly sensitive to modeling errors. In contrast, direct sensing of contact forces at the feet provides more accurate and immediate feedback. Beyond force magnitude, tactile sensing also enables richer contact interpretation, such as detecting force direction and surface properties. In this work, we show how tactile sensor information can be used inside the feedback of the control loop to achieve compliance of legged robots during ground contact. The three main contributions are (i) a fast and computationally efficient 3D force reconstruction method tailored for spherical tactile sensors, (ii) a tactile admittance controller that adjusts leg motions to achieve the desired GRFs and compliance, and (iii) experimental validation on a quadruped robot, demonstrating enhanced load distribution and balance during external perturbations and locomotion. The results show that the peak ground reaction forces were reduced by 55% while balancing on a beam. During a locomotion scenario involving sudden touchdown after a fall, the tactile admittance controller reduced oscillations and regained stability compared to proportional–derivative (PD) control.

20 January 2025 in IEEE Sensors Letters
Verified
Journal Article
Closed access
THThijs Van Hauwermeiren Van Hauwermeiren
Thijs Van Hauwermeiren Van Hauwermeiren
ASAnatolii Sianov
Anatolii Sianov
ACAnnelies Coene
Annelies Coene
GCGuillaume Crevecoeur
Guillaume Crevecoeur
This letter introduces Tweelie: a tactile wheel-shaped sensor with soft elastomer skin based on barometric pressure transducers. Tweelie enables high impact force reconstruction and localization over a curved spherical surface. Multiple contacts occurring simultaneously can be detected and inferred over a 6457 $\text{mm}^{2}$ surface. Based on the spatial distribution of 48 micro-electromechanical system (MEMS) sensors along a cylindrical surface, a graph is constructed to infer the contact state. The 3-D force localization is done by mapping the pressure readings onto an appropriate pressure distribution based on the shape of the Tweelie sensor; the 3-D force is obtained by integrating this distribution. Results show a localization error of 2$^{\circ }$ and regression error of less than 1 N for a single contact on rigid surface, enabling direct force localization and reconstruction for locomotion and other tactile applications.

11 August 2025 in Journal of Intelligent & Robotic Systems
Verified
Journal Article
Open access
PDF file
BGBabak Mehdizadeh Gavgani
Babak Mehdizadeh Gavgani
THThijs Van Hauwermeiren Van Hauwermeiren
Thijs Van Hauwermeiren Van Hauwermeiren
AFArash Farnam
Arash Farnam
MGMajid Ghorbani
Majid Ghorbani
Jeroen De KooningJeroen De Kooning
Jeroen De Kooning
GCGuillaume Crevecoeur
Guillaume Crevecoeur
This paper addresses the challenge of adapting the trajectory of robotic arms performing tosses of objects in dynamic conditions, where target locations can shift during operation. The high-speed, continuous nature of the tossing task necessitates precise and efficient adjustments to ensure that the objects reach their target location. Although nonlinear model predictive controllers (MPCs) are a valuable tool to calculate trajectories, it is difficult to assess in advance how long the calculations will take. They are, hence, difficult to use in tossing under dynamic conditions. To achieve accurate tossing under varying target locations within restricted computational time, we propose a soft switching multiple MPC strategy that enables smooth trajectory updates, while adhering to system constraints. By switching between multiple linear model predictive controllers, our approach ensures rapid trajectory generation, enabling seamless operation in unpredictable conditions. Experimental validation was conducted using a 6-degrees-of-freedom industrial robot arm, demonstrating the effectiveness of the proposed method in reducing computational load by a factor of 1300 compared to a benchmark based on nonlinear MPC. These results confirm its effectiveness for high-speed robotic tossing and suggest broader applicability in tasks requiring real-time trajectory adaptation.

3 March 2022 in TechRxiv
Verified
Preprint
Open access
PDF file
SWSander De Witte
Sander De Witte
THThijs Van Hauwermeiren Van Hauwermeiren
Thijs Van Hauwermeiren Van Hauwermeiren
TLTom Lefebvre
Tom Lefebvre
GCGuillaume Crevecoeur
Guillaume Crevecoeur
Cooperative multi-agent manipulation systems allow to extend on the manipulative limitations of individual agents, increasing the complexity of the manipulation tasks the ensemble can handle. Controlling such a system requires meticulous planning of subsequent subtasks, queried to the individual agents, in order to execute the master task successfully. Real-time planning is essential to ensure the task can still be achieved when subtasks execution suffers from uncertainty or when the master task changes intermittently requiring real-time reconfiguration of the plan. In this work we develop a supervisory control architecture tailored to the cooperation of two robotic manipulators equipped with standard pick-and-place facilities in the plane. We consider a toy task description where we control the planar position and orientation of an object. A time-invariant policy function is trained using deep reinforcement learning, which can determine a finite sequence pick-and-place maneuvers to manipulate the object into its desired configuration. A comparison is made between two strategies, with the distinction made based on different treatments of the final step. The more information is given to the policy the easier it trains. In return, it becomes less adaptable and loses some of its generalizability.