UST030

84 Focus | Lidar sense & avoid pattern, for example to scan a particular quadrant or direct beams at different angles. The phase of the light in the waveguide depends on the refractive index of its material, the length of the channel and the light’s wavelength. The index can be varied in several ways, most commonly by heating up the channel. This thermal tuning uses a micro- heater, a heating element built around each waveguide. The more current that is applied to the heater, the higher the phase change. In the lab there is no hysteresis effect, and as the waveguide cools the phase returns to the original value. However, this still needs extensive reliability testing in order to be used in production vehicles. It also adds significant complexity to the design of the optical beam-steering array, with thousands of heaters being needed across it, all with separate control lines. There are ways to simplify this, as the phase relationships are fixed, for example with alternating waveguides using the same phase change or a regular pattern of changes. This allows the control lines to be multiplexed together to supply particular channels. Other approaches seek to eliminate the need for power being consumed by the micro-heaters by using electro-optic materials that are compatible with silicon and silicon nitride. This is important, as the most common electro-optic material – lithium nitride – contaminates other tools used in the manufacturing process. Instead a liquid crystal can be used to induce a different index. This index change may not even be necessary. 2D scanning can be achieved with wavelength or phase changes, depending on the antenna array’s design. 3D scans require both techniques. Pure wavelength steering puts a lot of constraints on the laser source, as it needs to be tuned very accurately, while with a pure phase shift approach the source is fixed but consumes more power if thermal heating is used. With an all-phase shift approach the complexity of the actuator network is more complex than pure wavelength. The phase of the outgoing laser light also needs to be monitored, so a phase calibration scheme is needed on the chip. This monitors the waveguides and confirms the phase in real time by recovering the phase from adjacent antennas using an interferometer. This data can be fed back to the phase shifter controller to keep the beam on track. There are several options for the receiver in an optical beam-steering architecture. A lens can be used to collect light for the photo detector, using a similar approach to the mirror-based systems. However, the optical phase array can also be used to collect the beam. This requires additional components to isolate the received light from the transmitted light and redirect it to the photodetector. The advantage with this approach is to reduce the complexity of the optics, and it also allows the receiver to be implemented on the same chip using a silicon germanium process, significantly reducing the losses in the signal chain and the size of the Lidar sensor. Beam-steering architectures can use either of the receiver techniques, but an approach known as frequency modulated continuous wave (FMCW) is gaining traction as it can use standard photodetectors that are easier to implement on-chip. The modulation of the wavelength gives the timing of the return using interferometry: when combined with the original laser beam, the beat frequency corresponds to the distance. That means the receiving electronics don’t need February/March 2020 | Unmanned Systems Technology A coaxial design and highly automated production has achieved a smaller, lower cost Lidar sensor for short-range applications such as creating a cocoon of sensing around a vehicle (Courtesy of Velodyne) Researchers have developed an optical beam- steering chip for a solid-state Lidar sensor (Courtesy of University of Yokohama)