Chalmers Open Digital Repository

Welcome to Chalmers Open Digital Repository!

Here you can find:

Student theses and papers
Digital special collections, such as Chalmers modellkammare
Selected project reports

Research publications, reports and dissertations can be found in research.chalmers.se

Communities in Chalmers ODR

Select a community to browse its collections.

Now showing 1 - 2 of 2

Recent Submissions

Cobot + Vision to remove plastic straps from pallets
(2026) Lokku, Roger; Gustafsson, Alexander
This thesis aims at assisting Volvo Trucks to automate the de-strapping of the incoming pallets on a conveyor line using a vision-based robotic solution. Volvo provides a UR10e cobot, a Photoneo 3D-scanner and Aurora Design Assistant (ADA) to carry out this thesis. In return, Volvo requires the answer to the question whether this robot, camera, and software configuration is satisfactory to get the job done. The solution, however, needs to be divided into two main parts - building the system architecture along with integration of the given equipments and development of methods for de-strapping of the pallets. This thesis proposes the system architecture along with integration of the given equipments and the three methods that have been developed for de-strapping of the pallets. Each method is a combination of the force mode functionality of the robot and the image processing functionality via ADA after image acquisition via the camera. A python script is used to act as the middle man between the robot’s controller and ADA, facilitating data exchange and logging of data. The methods will be judged on the ability of the camera to identify and localize the strap, the ability of the UR controller to form a valid robot path to grab the straps and the cycle time. The accuracy and dependability of the robot, camera and software is judged by doing 30 cycles of grabbing the straps at various pallet offset and/or rotated from/at a fixed point. This will give quantitative data. Thus, answering Volvo’s question of "Can a UR10e cobot, a Photoneo 3D-scanner and ADA be used to de-strap their pallets?". The calibration error of the estimated homogeneous transformation matrix was found to be 3.5 mm. The fastest method was Method 3 with an average cycle time of 39.24 s. The camera, when deployed on the production line, was able to correctly determine the presence of straps with a success rate of 76.79% and the success rate of the colour identification of the straps was 75.00%. The camera was found to be able to solve the strap identification and colour identification but having a 3D camera is deemed excessive, and the image quality could be the limitation of the robustness of strap identification and colour identification.
Robotic Phantom Knee and Digital Twin Platform for Early-Stage Exoskeleton Validation: Nonlinear Stiffness and Tuneable Soft-Tissue Compliance
(2026) Habrah, Wajih
Testing wearable exoskeletons and rehabilitation devices directly on human subjects is often expensive, time-consuming, and ethically challenging, especially during early-stage development when designs and control strategies are frequently updated. In addition, human trials introduce variability that complicates systematic benchmarking and controller verification. This thesis presents the design and development of a robotic phantom knee (RPK) testbed and a real-time Unity-based digital twin (DT) intended to provide a controlled bench-top platform for evaluating exoskeleton controllers. The developed system combines a single-degree-of-freedom knee hinge with antagonistic actuation using two Koala BEAR linear actuators coupled through an elastic linkage to shape the joint’s passive resistance. To emulate tuneable soft-tissue and interface compliance, the platform integrates inflatable chambers whose stiffness is adjusted via closed-loop pressure regulation using solenoid valves, a pump and reservoir, and pressure sensing. A modular ROS 2 (Jazzy) software stack on a Raspberry Pi 5 handles hardware communication, pressure control, and safety functions (including E-STOP and release-air). A Unity-based DT connects via ROS–TCP to visualize system state in real time and support interactive parameter tuning and experiment execution. Subsystem validation demonstrated stable ROS–TCP streaming, safety handling (ESTOP and release-air), and closed-loop pressure tracking suitable for bench-top trials. The platform supports systematic controller development and validation before human testing by enabling controlled experiments under configurable compliance conditions.
Arcturus – Design och konstruktion av en vektorstyrd modellraket
(2026) Bergstedt, Frode; Eklind, Tilde; Enell, Anton; Gustafsson, Victor; Randerz, Viggo; Richardt, Christian
The goal of the project was to design and build a thrust vector controlled (TVC) model rocket, with an aim to stabilize the rocket with active control. The aim was split into three parts: (1) simulation; (2) static test with free rotation; and (3) free launch. To comply with Swedish laws, the rocket was restricted to maximum weight of 500 g, with a maximum 100 g reserved for propellant. With these design limitations in mind, the rocket body and systems for both the electronics, TVC, and parachute were custom made, with the goal to minimize weight. The designed rocket weighed a total of 499,2 g with propellant. However, parts like the custom flight computer did not work as intended, thus a new makeshift solution was required for launch. After the rocket was designed and built, three static test launches and one live flight were executed. Even though the simulations gave promising results, none of the launches proved successful. This could be due to a discrepancy between simulation and reality. In the end, the goal was achieved, but the aim failed on all three parts, which could have been remedied by a longer research and development phase.
Model-Based Estimation of Clamp Force in Bolted-Joint Tightening
(2026) Tallanje, Athreya
Bolted joints are widely used in mechanical assemblies, and their reliability depends strongly on the clamp force generated during tightening. Direct measurement of clamp force in industrial applications is often impractical. Conventional torquebased tightening methods estimate clamp force by assuming a constant coefficient of friction (CoF). However, the accuracy of this assumption is influenced by surface conditions, lubrication, tightening speed, and other factors. Therefore, accurate clamp force estimation remains a challenging problem. Previous Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF)-based approaches achieved estimation errors of approximately 3–5%, but were found to be sensitive to initial state assumptions. In addition, the coupling between clamp force and CoF can produce similar torque outputs for different state values, resulting in non-unique estimates. This thesis focuses on developing system formulations for tightening in the elastic region. The main states considered are clamp force and CoF, which are strongly coupled through the torque expression. To analyze the resulting non-uniqueness in state estimation, both local and global observability perspectives are considered. Multiple observer-based estimation approaches are then evaluated, including an LMI-based Luenberger observer using an LPV formulation, an output-inverted EKF, and a Rao-Blackwellized Particle Filter (RBPF). The EKF and RBPF formulations reduce the dimensionality of the directly estimated state space. The results indicate that the system is theoretically observable from the local observability analysis, but the near-marginal values of the observability matrix suggest limitations in state estimation. The output-trajectory-based observability study further shows trajectory-dependent regions of practical unobservability, making it difficult to generalize a single unobservable region across different tightenings. The LMI-based Luenberger observer was highly sensitive to initial-state errors and relied strongly on the motion model, whereas the output-inverted EKF was mainly limited by derivative-induced noise amplification in the measurement model. The RBPF provided improved estimations by maintaining multiple hypotheses, but its performance depended strongly on both initialization and the nonlinearity of the output trajectory. Overall, the study shows that reliable reconstruction of clamp force and CoF is fundamentally constrained by weak observability, low effective CoF contribution to the measured output, and model mismatch under nonlinear tightening conditions.
Inverse kinematic optimization and MPC on a four-link hydraulic manipulator
(2026) Billing, Konrad; Djäknegren, Filip
Operating individual joints on a hydraulic manipulator can be tedious and unproductive with limited experience. Therefore, this thesis presents a method to ease operator strain and lessen experience requirements by enabling cartesian control of the robot tool. This is achieved with differential closed-loop inverse kinematics and linear model predictive motion control. References are generated with a pseudo-inverse Jacobian algorithm formulated as a quadratic program with a secondary null-space objective to exploit redundant joints. A sensitivity analysis of a non-linear hydraulic model resulted in a linear first-order model which was used to design an MPC that tracks the references. Results from a simulated robot show robust tracking performance even when the robot is exposed to noise and unmodeled disturbances. The designed null-space objective manages to control null-space motion and the CLIK is shown to reject unmodeled disturbances.