A team of researchers at the University of Berkeley has developed new state-of-the-art laser technology which should lead to a better price-quality ratio for driverless vehicles.
Self-driving cars are increasingly being seen as the means of transport of the future, or at least as an alternative that will sooner or later win out over cars with human drivers. By 2018, autonomous cars should be on our roads with mass adoption predicted for 2050. However, there still remains a number of obstacles to be surmounted – legislative updates and changes in people’s mindset required – before we reach that stage. On a more down-to-earth level, the price of self-driving cars is also currently another obstacle to mass adoption. A key aspect of autonomous road vehicles is lidar (a composite of ‘light’ and ‘radar’), a remote sensing technology that measures distance by hitting a target with a laser and analysing the reflected light. This enables the driverless car to analyse its environment and adapt its movements accordingly. However the main disadvantage is that the technology is both cumbersome and expensive. First generation Google Cars are fitted with a lidar that costs $80,000. Meanwhile progress is being made in this field. German firm SICK now makes lidar scanners for just $10,000. However, a team of engineers at the University of California, Berkeley, has just published its findings on a system which could substantially improve the price-quality ratio of this vital piece of equipment.
Lighter, smaller, cheaper and less power-hungry
A lidar works by repeatedly changing the wavelength of a laser, so that the sensor can accurately identify the light when it bounces off an object in the vicinity and returns to the sensor. This enables it to calculate the distance of the object from the vehicle. These wavelength changes require precise manipulation of a mirror, or sometimes multiple mirrors. With the traditional approach, a separate electrical device is used to slide the mirrors to and fro. But the UC Berkeley team has developed a system whereby the mirrors are adjusted by the laser itself, which they describe as a ‘self-sweeping laser’. Doing away with the external electrical device means that the power consumption, size, weight and cost of the lidar equipment can all be drastically reduced. This innovation should in turn help to make the price of autonomous cars far more affordable, accelerating general adoption of driverless vehicles.
Potential major impact on the robotics sector
But the possibilities created by the new approach do not end there. This pioneering technique could also open the way to making smaller vehicles – which are not suited to carrying the cumbersome lidars currently on the market – self-driving. Perhaps it might soon be the norm to move around on the golf course in a self-driving golf cart. Recently Weijian Yang, lead author on the UC Berkeley study, told technology magazine Wired that this work could also enable new approaches in medical imaging and in robotics, especially for unmanned aerial vehicles. The deployment of fleets of autonomous drones that can make deliveries, perform rescue operations, carry out building repairs and irrigate crops is not going to happen tomorrow, but the potential advances arising from the new lidar technology certainly make it worthwhile investigating this approach further. Meanwhile Google Cars seem to have won over large numbers of people in Austin, Texas.