Internship: Enhancing a Probabilistic LiDAR MOT Framework

Are you looking for an internship or graduation project in unmanned vehicles?

About Demcon

DEMCON Unmanned Systems develops unmanned surface vessels (USVs) with increasing levels of autonomy. An important step in autonomous navigation is the detection and classification of other vessels and other objects in the water, in order to respond accordingly. The USV is equipped with four cameras for 360-degree coverage that can provide input for any classification algorithm.

What will you do?

We are seeking an intern to contribute to the enhancement of a probabilistic LiDAR-based Multi-Object Tracking (MOT) framework. The framework combines a detector and tracker using a point-based approach and requires improvements in the following areas:

• Clustering Algorithm Improvement
  • Replace the current Euclidean clustering (PCL library) with an adjustable kernel-based approach that adapts to object locations.
  • Aim to reduce oversegmentation and ensure the correct number of clusters matches the true number of objects (e.g., vessels).
• Pointcloud Index Tracking for Occupancy Mapping
  • Develop a method to map downsampled points back to their original indices.
  • Propagate cluster assignments from the downsampled cloud to the raw points, enabling precise retrieval of original indices for each cluster.
• Shape Estimation Optimization
  • Investigate alternatives to the existing PCA-based shape estimation.
  • Implement an adjustable algorithm tailored to the spatial distribution of individual clusters.
• Debugging and Automation
  • Resolve ROS 2 message overflow issues to enable the MOT framework to run across multiple scenarios without manual supervision.

What do you bring?

• Proficient in C++ and Python.
• Experience with ROS 2 is a plus.

This role offers hands-on experience with advanced LiDAR data processing, robotics, and multi-object tracking systems.

What do we offer you?

DEMCON offers a dynamic, growing and innovative organization. With our complex projects and multidisciplinary teams we offer a good climate for this graduation project/internship. Thanks to enthusiastic and dedicated colleagues, our working atmosphere is good, open and informal. You will be supervised (daily) by a colleague from our department. We welcome own initiative and give you the space to express this. Furthermore, we offer you a graduation allowance and a lot of flexibility.

Please reach out to Hessel Edelijn (Hessel.edelijn@demcon.com) if you have any questions