Over the past couple of years, the University of Michigan has established Mcity, a large open-air test lab that mimics urban street scenarios with fake buildings and 8 kilometres (5 miles) of urban roads, including roundabouts, traffic light intersections, a tunnel, construction sites, a bridge and streets with a diversity of surfaces. As the first OEM, Ford now has started testing at Mcity. The test rides are carried out with a modified Focus sedan that is equipped with the sensors necessary to orient itself in the traffic. Among other sensors, a LIDAR infrared scanner creates a virtual 3D map of the surroundings at distances up to 60 metres.
A spokesperson of Ford Germany disagreed with a report from online magazine Wired stating Ford was pursuing a strategy for autonomous driving that resembles Google’s approach of driving entirely autonomously, aiming to avoid situations in which the car realises that the situation gets too complex for the algorithms to handle hand therefore calls the driver back to duty. Such a phase indeed is critical, since OEMs calculate it might last about ten seconds – a time span during which the vehicle continues to drive at potentially rather high speed through a situation it just realised that is difficult. And indeed, the handover from the car to the vehicle after terminating the phase of driving autonomously for which ever reasons is something that puzzles the design engineers.
However, it is not Ford’s strategy to skip this step in the evolution of autonomous driving, the spokesperson stressed. “Our development strategy is in line with what other large OEMs are doing”, he said.
Ford’s test vehicle is however using “Driver-in-Control” analysis algorithms developed by the US carmaker. These algorithms have previously been tested in Ford’s VIRTTEX driving simulator. VIRTTEX indeed focuses on driver-vehicle interaction topics, the goal of the research with this simulator is establishing what Ford calls a holistic driving experience. In collaboration with