USBV-Inspektor

My Role in this project

I was a student assistant working on the simultaneous localization and mapping (SLAM) of the sensor suite using data from laser scanner, camera and IMU. Together with my colleagues, I conducted experiments (e.g. in an abandoned metro station or other public places) during which we used the self-built sensor suite to record so-called “rosbags”. In addition, I implemented parts of our SLAM algorithm using C++ and the Robot Operating System (ROS).

Motivation

In public places, train stations and on trains, forgotten luggage is part of everyday life. Even if the bags are harmless in most of these cases, they must be treated as possible sources of danger until the situation has been clarified. If the special forces determine that it is indeed an improvised explosive device (IED), all necessary measures are initiated immediately. Thus, it is important to minimize the risk for the emergency services and to secure as much evidence as possible to identify the perpetrator.

Objectives and procedure

The project therefore deals with the research and development of a multimodal sensor system (sensor suite), consisting of a millimeter wave scanner, a 3D environmental perception system and a high-resolution digital camera. These are integrated in a housing and mounted on a remote-controlled robot platform. The millimeter wave scanner is used to x-ray and inspect the inside of the luggage. The detection of the surroundings and the camera are used to investigate and digitally secure evidence at the crime scene as well as to survey the crime scene. The data are linked, visualized and saved for later legal evaluation.

Innovations and perspectives

In the future, the sensor system will enable the emergency services to quickly identify explosive devices or incendiary devices in luggage without having to expose themselves to the direct risk of explosion. In addition, the sensor suite can be used to permanently secure reliable evidence at the crime scene, as it is captured and digitized in three dimensions and images, and is therefore available for legal assessment at any time.

Raphael Voges
Raphael Voges
Doctor in Robotics

My research interests include error modeling, sensor fusion, SLAM, state estimation and AI in the context of autonomous driving.

Related