Multi-Sensor Data Fusion

Our research focusses on developing the theoretical and practical techniques and methodologies pertaining to Multi-Sensor Data Fusion, with a specific focus on applications with a requirement for high precision situational awareness in a multi-target, multi-source environment. Multi-Sensor Data Fusion is a cross domain, enabling technology with applications across a variety of domains, however the historic focus of our research has been in the automotive, robotics and defense and aerospace domains.

Research Topic Description

The principal aim of Multi-Sensor Data Fusion is to combine sensory information from disparate sources in order to improve knowledge about the state of an observed object. Developments in digital technology have resulted in a proliferation of sensors with significant variations in physical size and capability; an increasing speed and availability of wireless communications; and significantly increased processing capacity. While this has realized a number of opportunities across different domains, it has also introduced a number of challenges, particularly when the registration and noise characteristics of such sensors are poorly understood. The goal of the Multi-Sensor Data Fusion group at fortiss is to facilitate the integration of both dedicated and disparate sensors using System of Systems principles. Specifically, this means that we integrate independent task oriented systems into a single complex system with improved functionality.

Our research focusses on developing the theoretical and practical techniques and methodologies pertaining to Multi-Sensor Data Fusion, specifically the use of software to enable effective integration of sensor systems. Our approach is multi-disciplinary in that we consider all aspects of the multi-sensor data fusion problem. This includes the high performance signal and image processing algorithms, fusion/communications architectures, sensor/data fusion algorithms and application specific requirements. While we have a specific interest in the application of machine learning and artificial intelligence techniques (e.g. Swarm Intelligence, Intelligent Agents and Artificial Neural Networks), we also pursue research into the more traditional methods for state estimation (e.g. Probabilistic Graphical models, PHD filters).

Multi-Sensor Data Fusion can be considered as an underpinning technology whose functionality is used to solve specific applications. Therefore, the specific applications of our research are cross domain, with the historical focus of our researchers being the automotive, robotics and the aerospace, defense and security domains. However, we have an interest in all applications where the integration of sensory information in a system of systems architecture can be used to solve interesting applied problems.


Daniel Clarke – Group Leader

Feihu Zhang – Staff Researcher

Selected Projects

Imagery Fusion Investigation