Unmanned Systems Technology 010 | nuTonomy driverless taxi | Embedded computing | HFE International marine powertrain | Space vehicles | Performance monitoring | Commercial UAV Show Asia report

8 Platform one The increasing popularity of Lidar laser detection sensors in driverless cars and UAVs is leading sensor makers to develop smaller and more cost-effective devices (writes Nick Flaherty). Velodyne for example has launched a higher resolution version of its 16-channel Puck Lidar for more detailed 3D imaging. Puck Hi-Res keeps the VLP-16 Puck’s 360° horizontal field of view (FOV) and 100 m range, but delivers a 20° vertical FOV for a tighter channel distribution – 1.33° between channels instead of 2.00° – to deliver greater detail in the 3D image at a longer range than 100 m, depending on the size of the object being detected. That will enable the host system not only to detect but also better discern objects at these greater distances. Puck Hi-Res will compete with a new solid-state Lidar from Quanergy Systems, the Q3, that the company claims will be available for $250. Quanergy has raised $90m in a round of funding to develop the technology and ramp up production. The sensor uses standard semiconductor manufacturing processes and has no moving parts, thereby offering far lower cost, higher reliability, superior performance, increased capability, smaller size and lower weight compared to traditional mechanical sensors. The company currently has pre- production contracts with multiple global customers for the sensor. “Innovation in Lidar technology represents one of the largest opportunities unfolding around the globe, and the funding will enable us to accelerate development, scale faster and expand our engineering team,” said Dr Louay Eldada, CEO of Quanergy. Investors include automotive component supplier Delphi as well as chip maker Samsung, giving an indicator of the likely customers and technology partners. Startup Scanse, in California, is developing a different kind of Lidar that will also enable a $250 unit. Its system relies on a new sensor, developed by a company called PulsedLight that was acquired by GPS specialist Garmin in January 2016, which uses a time-of-flight ranging. It involves sending out laser light that consist of a series of micro-pulses and measuring the difference in phase of the returning light. It can use less power than existing Lidar designs, but has a shorter range of around 40 m and a lower update rate of 500 Hz rather than 2-10 kHz. That makes it more suitable for obstacle detection for UAVs, as opposed to environment scanning for mapping in driverless cars. Elsewhere, work continues to integrate solid-state lasers into small Lidar systems. For example, researchers at the Massachusetts Institute of Technology (MIT), working with the US DARPA research agency, have developed Lidar beam steering receivers that can be made on the same silicon wafers as modern processor chips. The Lidar chips are produced on 300 mm wafers, making their potential production cost of the order of only about $10 each at volumes of millions of units per year. The non-mechanical beam steering chip measures 0.5 x 6 mm, and would be combined with a solid-state laser for a lightweight, low-cost sensor that can respond 1000 times faster than mechanical sensors, according to Chris Poulton, a researcher in the Photonics Microsystems group at MIT. The current steering range of the beam is about 51°, being limited by the spacing between the antennas. Reducing this spacing becomes challenging because there is a limit to how small silicon waveguides can be while still confining light adequately, although the technology should support near 100° steering, said Poulton. Even with a limited steering range, the lower cost means multiple sensors on a vehicle would provide full 360° image.  Rather than using direct time-of-flight, the sensor uses a coherent light method where the system reacts only to the originally transmitted light that is received back by a phased array antenna. This reduces the effect of sunlight, which can be a large noise factor in Lidar systems, and allows for modest photodetectors instead of expensive avalanche photodetectors or photo- multiplier tubes, which are expensive to integrate in a silicon photonics platform.  At the moment, the on-chip Lidar system can detect objects at ranges of up to 2 m, with plans to reach 10 m in 2017. But there is a clear development path towards ‘Lidar on a chip’ technology that can reach 100 m, with the possibility of going even further, said Poulton by using materials such as silicon nitride and a larger phased array antenna. DARPA has recently created a follow- up programme called Modular Optical Aperture Building Blocks to develop the technology further. Lidar advances gather pace Sensors October/November 2016 | Unmanned Systems Technology Researchers at MIT have developed a $10 Lidar sensor that is only 6 mm long (Courtesy of Chris Poulton, MIT)