The Unmanned Port Security Vessel: An autonomous platform for monitoring ports and harbors

This paper describes the development of the Unmanned Port Security Vessel (UPSV), a small autonomous surface vehicle designed to support maritime domain awareness in port and harbor environments. The UPSV is capable of rapidly producing fine resolution, shallow-water bathymetry maps using a multibeam sonar, detecting chemical threats using an on board mass spectrometer and monitoring oceanographic parameters using off-the-shelf instruments. In this paper we discuss the significance and scientific advantages of sensor fusion on a small autonomous robotic platform, as well as detailing the control methods chosen for such a vehicle. Command and control of the robotic platform is implemented using the lightweight communications and marshalling (LCM) library, a general purpose software infrastructure for sensor integration and feedback control. Based on this foundation we have built a flexible software architecture that includes mission planning, real-time navigation and control, payload management and telemetry to support a variety of maritime domain awareness missions. Using the developed software and control methods, the UPSV is capable of autonomously generating high fidelity bathymetric maps while simultaneously conducting chemical surveys with a mass spectrometer. We demonstrate the utility of the software architecture and hardware integration through results from both simulation and experiments. We present the results of a set of field trials to evaluate the performance of the UPSV in operational settings. These experiments include system identification for estimation and control, feedback control performance evaluation, demonstration of autonomous high-resolution bathymetric mapping and demonstration of in-situ chemical sensing. The performance of the UPSV´s control system is characterized with the presented navigation data, and the results of the bathymetric surveys and in-situ chemical sensing are detailed. The implementation of this functionality on a small,- deployable platform is discussed in the context of similar projects on unmanned surface vehicles.