Development of Sensor Module for Terrain Mapping in Forest Environment

Abstract

Unmanned ground vehicles (UGVs) operating in uneven and unstructured terrain are often limited by fixed sensor mounting, which can reduce mapping quality, situational awareness, and sensor performance when the vehicle body experiences pitch, roll, and vertical motion. This thesis presents the development of a compact four-degree-of-freedom (4-DOF) gimbal prototype for a terrain-mapping module on a quadruped UGV. The system provides yaw, pitch, roll, and vertical elevation, enabling both self-stabilization and adjustable sensor positioning.

A prototype was developed through an iterative process involving CAD design, 3D printing, assembly, and integration of motors, sensors, and control hardware. The final concept included a screw-driven elevation mechanism, a yaw actuator, and a pitch-roll stabilization structure with IMU-based control. A Raspberry Pi 5 served as the central processing unit, together with a servo driver, a stepper motor driver, and an IMU for angle estimation.

To support stabilization, IMU filtering and control strategies were implemented and tested experimentally. A complementary filter and a Kalman filter were compared for angle estimation, and the Kalman filter provided the most suitable performance in terms of drift reduction and signal quality for this application. For pitch and roll control, a PD controller and two active disturbance rejection control (ADRC) approaches, referred to as Simple ADRC and Hybrid ADRC, were evaluated. The PD controller was not sufficiently responsive to fast tilt disturbances, while the ADRC-based methods provided better stabilization performance. Simple ADRC performed best on the roll axis, whereas Hybrid ADRC gave the best results on the pitch axis.

In addition to self-stabilization, the prototype demonstrated yaw rotation and vertical elevation. The yaw function achieved commanded rotations with an accuracy of approximately plus/minus 2 degrees, while the elevation mechanism raised the system from about 24 cm to 33 cm above the UGV platform in 1 minute and 27 seconds. A simulation study in Adams/View was also carried out to investigate passive vibration mitigation. For three representative walking cases, the simulated spring-damper support reduced transmitted motion in heave, pitch, and roll, suggesting that passive isolation could complement active stabilization and improve performance.

Overall, the thesis demonstrates the feasibility of a compact 4-DOF gimbal system for a quadruped UGV sensor module. The result is a functioning proof-of-concept prototype that combines mechanical design, embedded control, and vibration mitigation in one integrated system. Although limitations remain in structural robustness, packaging, and full system integration, the findings provide a solid basis for further development of a more robust sensor platform.

Popular Abstract

A Gimbal That Helps Robot Sensors Stay Steady on Rough Terrain

Robots that move through forests, fields, or other rough environments often struggle to collect clear sensor data because their bodies tilt and bounce. In this thesis, a compact sensor module was developed for a quadruped unmanned ground vehicle and can move in four ways: rotation around yaw, pitch, roll, and vertical elevation. The goal was to improve terrain mapping by keeping the sensor package more stable and better oriented toward its surroundings.

The project was developed step by step through CAD design, 3D printing, mechanical assembly, and integration of motors, sensors, and control electronics. The final prototype used a screw-driven lifting mechanism, a yaw motor, and a pitch-roll stabilization structure controlled with feedback from an inertial measurement unit (IMU), which measures motion and orientation. Different methods for interpreting IMU data were tested, and the Kalman filter provided the most useful signal quality for this application.

Several control strategies were also evaluated to stabilize the system during motion. The results showed that the ADRC-based methods handled fast tilting disturbances better than a simple PD controller. In practical tests, the yaw mechanism reached commanded angles with an accuracy of about plus or minus 2 degrees, and the elevation mechanism raised the sensor module from about 24 to 33 centimeters above the robot platform in 1 minute and 27 seconds.
In addition to the physical prototype, simulations were carried out to study passive vibration damping. These showed that a spring-damper support could reduce transmitted heave, pitch, and roll motion in several walking cases, suggesting that passive damping could complement active stabilization. Overall, the project shows that a compact stabilized sensor platform for a quadruped robot is feasible and provides a solid basis for further development of a more robust terrain-mapping system.

In the future, this kind of technology could help robots collect clearer data and move more safely in places where the ground is uneven and the terrain is hard to read.

• author: Hellquist Tovar, Vincent and Gruber, Konrad
• supervisor: Aylin Ahadi
• organization: Mechanics
• course: FMEM01 20261
• year: 2026
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• keywords: 4-DOF gimbal Quadruped UGV Sensor stabilization Terrain mapping IMU-based control Vibration mitigation analysis
• other publication id: LUTFD2/(TFME-26)/5049-SE(1-67)
• language: English
• id: 9234102
• date added to LUP: 2026-06-12 10:31:55
• date last changed: 2026-06-12 10:31:55