Unmanned Systems Technology 022 | XOcean XO-450 l Radar systems l Space vehicles insight l Small Robot l BMPower FCPS l Prismatic HALE UAV l InterDrone 2018 show report l UpVision l Navigation systems

36 Focus | Radar systems automotive collision-avoidance market. These programmable radar-ready integrated circuits include an FMCW waveform generator and six-channel intermediate-frequency processor with a multiplexed analogue-to-digital converter (ADC). These are used with a low-gain transmit antenna to illuminate nearby objects, and echo signals are simultaneously received from an array of five low-gain receive antennas with no scanning involved. An object’s range is derived from the echo signal’s beat frequency, while its Doppler frequency, which gives the radial velocity, is obtained from echo phase variation over sequential observations. Its azimuth and elevation position relative to the UAV is determined by comparing the phases of the signals captured by the five receive antennas. However, tests of these systems shows that the radar is only needed to detect objects in the 300-800 m range from the original value of 1 nautical mile (1852 m), allowing a smaller, lower-power design. The initial design needed five receive antennas to unambiguously provide azimuth and elevation position knowledge for each detected target at a 10 Hz refresh rate. Analysis of angular rates of change for targets in the 300- 800 m detection range (assuming speeds commensurate with small UAVs) indicates that these angle update rates can be relaxed (from 10 to 5 Hz) while maintaining the required accuracy in estimated target position. This reduces the requirements on the real-time signal processor in terms of the number of parallel channels requiring simultaneous processing, from five to two. By simultaneously processing two channels during one 100 ms observation period, the target’s elevation angle can be updated, while in the next 100 ms period the target’s azimuth angle can be updated. That enables all the FPGA processing to be performed in a single low-cost device with no need for off-chip memory resources, and still use the same FMCW radars in the automotive collision avoidance market. While the current automotive collision detection specification is clear, high- definition radar imaging systems in driverless cars are facing the same specification issues as airborne platforms, with the requirements being unclear. These systems are intended to challenge the more expensive and power-hungry laser-based ranging (Lidar) sensors in building up detailed maps of the environment at similar resolutions around a vehicle to prevent collisions. Having a single sensor technology rather than a combination of Lidar and radar is seen as reducing the complexity of the sensor fusion algorithms and subsystem procurement. Metamaterials One key emerging technology for radar in both airborne and high-definition radar imaging systems is to use metamaterials for the antenna at the front end. Metamaterials have been engineered to respond to electromagnetic signals in ways that are not found in nature. They are usually built up of repeating patterns, at scales that are smaller than the wavelengths of the phenomena they influence. For example, in a radar system, the scale of the patterns will depend on whether the radar is operating at 24 or 77 GHz. Applying a series of voltages across the elements in the metamaterial allow the permittivity to be changed, altering the phase of the signal as it travels through the antenna. Using metamaterials can produce smaller radar systems with lower power consumption, as far less computing power is required and more functions such as beam steering can be handled without moving parts instead of using mechanical gimbals. Beam scanning A key advantage of using a metamaterial is the ability to change the phase of a signal as it moves through the antenna, on both the transmit and receive sides. Changing the phase via the voltage across the metamaterial October/November 2018 | Unmanned Systems Technology Some radar systems can be used for collision detection as well as high-definition imaging in unmanned ground systems (Courtesy of Arbe Robotics)