A Norwegian carrier's launch of a pilotless container ship next year could lead the way to a new era in sea transportation and shipping if successful.
Norway-based Yara International ASA, a self-touted "Tesla of the seas," claims the eletric-powered vessel will be the world's first autonomous commercial ship in operation when it sets sail next year to begin shipping fertilizer between ports in Norway.
Kongsberg Gruppen, a merchant marine, offshore, oil and gas industry, defense and aerospace systems provider; will design and produce the ship. This will include integrating the sensors, onboard computers, and mapping software for the autonomously piloted commercial vessel. Kongsberg Gruppen will also integrate the ship's electric drive, battery, and propulsion control systems.
Similar to driverless cars and trucks, driverless ships require advanced neural networking capabilities to process massive amounts of data to make piloting decisions. Yara's pilotless boat will thus rely on cameras, radar, and other sensors (whether the ship will use LiDAR remains undisclosed) to monitor sea and weather conditions, signages, fish and other wildlife, rocks, floating objects, and other boats. The sensors will send the data to a central machine-taught computer system, which will share the same underlying technology used to control driverless cars and trucks.
Automatic pilot features have been available for motor-powered and sailing vessels for decades. However, for commercial shipping, direct human intervention is often required to pilot boats in the event of extreme weather and wave conditions. Under international maritime law, a human pilot must be able to assume control of the ship in the event of an emergency, similar to how Level 2 and Level 3 driving for cars and trucks requires that human driver must be ready to take control of the car at all times.
The massive replacement commercial of commercial ships — which transports over 90% of all goods worldwide — with pilotless vessels would have an obvious effect on maritime pilot employment. It could also help to reduce the number of deaths and injuries in the commercial shipping sector since pilotless boats would require fewer crew members to operate.
Intel has introduced a USB-connected mobile development kit for deep learning applications that could be used to help offset expensive development costs for neural network systems for self-drive vehicles and other applications. The device's affordable price of $79 could help to lower the barrier of entry for researchers and electronics suppliers. Some developers, for example, might have the brainpower to develop advanced machine learning solutions for carmakers and other OEMs, but otherwise lack the financial resources to invest in expensive engineering lab equipment.