Situational awareness on inland waterways with multimodal, standalone sensor box: design and experiments
|
| Abstract | Dynamic environmental information obtained by perception sensors, together with proprioceptive sensor data, are a fundamental part of a vessel's control and state estimation system. This paper discusses the design and first experiments of a mobile, standalone, multimodal sensor box integrating both proprioceptive and perception sensors, that aims to provide convenient access to fused, synchronised sensor data. A second design criterion is to have external interfacing capabilities for extending existing vessel systems such as the Automatic identification system (AIS), RADAR, and Electronic Navigational Chart (ENC). As such, this work contributes to increasing the situational awareness on inland waterways, which is seen as an important first step towards safe, reliable control and fleet coordination, especially on smaller waterways where vessels are subject to close encounter manoeuvring. The box is designed for: (i) onboard vessel deployment, possibly interfacing the sensor box with a vessel's actuation system; and (ii) as a shoreside (sub)system to increase situational awareness in critical, high-traffic areas. Experiments focus on exploiting fused sensor data (from Light Detection and Ranging (LiDAR), Inertial Measurement Unit (IMU) and Global Navigation Satellite Systems (GNSS)) for mapping the environment, detecting relevant waterway features, and localisation of a vessel. The results are benchmarked against an existing, fully embedded, high-performant vessel sensor system. Hence, this paper provides insights in performance characteristics of a standalone sensor box, which proved to have matching performance to the fully integrated vessel sensor setup. |
|
