(c)iStock.com/Scharfsinn86

Ford has revealed details on how its self-driving vehicle – the Ford Fusion Hybrid – will manoeuvre itself on snow-covered roads in winter.

The company's winter weather road testing is undertaken in Michigan, including at Mcity – a 32-acre, real-world driving environment at the University of Michigan.

To operate in snowy conditions, Ford Fusion Hybrid autonomous vehicles first have to scan the environment and formulate high-resolution 3D digital maps. This is done in ideal weather conditions first where the vehicles develop accurate digital models of the road and surrounding infrastructure using four LiDAR scanners that generate a total of 2.8 million laser points a second. The vehicles generate so many laser points that some even bounce off falling snowflakes or raindrops. As a result, Ford developed an algorithm that recognises snow and rain, filtering them out of the car’s vision.

Ford autonomous vehicles collect and process a diverse set of data about the road and surrounding landmarks, collecting up to 600 gigabytes per hour – more mapping data than the average person uses in mobile-phone data in 10 years. The accuracy of vehicle navigation is higher than that of GPS and these vehicles are able to precisely locate themselves to within a centimetre.

Ford also makes use of cameras and radar to keep track of the environment around the vehicle, with the data gleaned from all sensors – including LIDAR – integrated in a process known as sensor fusion, bring about 360-degree situational awareness. This process ensures the inactivity of one sensor – probably caused by ice, snow, grime or debris buildup on a sensor lens – does not halt autonomous driving. Eventually, the cars will be able to deal with ice and grime buildup via self-cleaning or defogging measures.