16 August/September 2023 | Uncrewed Systems Technology The Maritime Research Institute Netherlands (MARIN) has combined its Dolphin simulation software with the Voyager AI-based intelligent autonomous navigation software from Robosys Automation to test the safety and performance of obstacle avoidance systems (writes Nick Flaherty). MARIN successfully trialled the original Voyager 100 reference software in its ship simulator as part of the Netherlands Joint Industries Project, which culminated in North Sea trials with three ships. The aim of the project was to measure the compliance of autonomous navigation software with international collision regulations. Voyager AI now integrates and controls multiple ships’ hardware, including vessel systems, switches, relays and monitoring systems and alarms. These can range from deck and hull sensors to creating an automatic response to initiate a fire extinguisher. “The original Voyager 100 performed well in simulation tests, so we are looking forward to evaluating the improvements included in Robosys Automation’s latest version,” said Dimitri van Heel, senior project manager for MARIN. The AI version has improved collision avoidance algorithms developed using simulation and digital twin technology to emulate a vessel in operation. Multivessel avoidance manoeuvre scenarios have also been enhanced. This version also includes a loss of comms recovery function, to give operators mission continuity in the event of a breakdown in comms. An integrated remote and autonomous platform alarm, and a monitoring and control system, provide full support for uncrewed craft. Vessel tracking and dynamic navigational path changes are made in accordance with a vessel’s manoeuvrability and its other performance characteristics, to produce a manoeuvring path matched to that vessel. Testing AI boosts collision tests MARIN and Robosys Automation have teamed up to test the safety of collision avoidance systems Researchers in France have developed a 256-channel beam steering device for Lidar sensors in driverless cars (writes Nick Flaherty). The development at CEA-Leti in Grenoble involved 3D Integration of through-silicon vias (TSVs) and flip-chip technologies to package the optical devices. The 10 µm-diameter TSVs significantly improve interconnect density by distributing them on the whole backside surface of the devices. Combining 40 µm fine-pitch flip chips on a silicon interposer, the integration increases the performance and compactness of Lidars, while lowering their cost for use in autonomous vehicles. The researchers developed steering concepts using wafer-level silicon technology, including optical phase arrays. A 1550 nm wavelength beam was then brought into the circuit through grating couplers, then guided through silicon oxide waveguides to the steering area. This provides precise, latency-free information on the position and speed of obstacles surrounding road transport vehicles, independently of light conditions and with sharp angular resolutions. Driverless cars Beam steering for Lidars
RkJQdWJsaXNoZXIy MjI2Mzk4