OFDM radio based range and direction sensor for robotics applications

One of the fundamental problems in building an autonomous robot is the perception and reconstruction of its environment, which is traditionally achieved with sensory systems that use camera, laser, sonar, or radar. This paper proposes the use of an Orthogonal Frequency Division Multiplexing (OFDM) based radio to sense the environment. Our solution uses an OFDM wireless transmitter, such as the one from Wi-Fi, on the robot to transmit a signal which is then reflected back by obstacles, and the reflected signal is captured with an onboard receiver. The receiver analyzes the incoming signal for multipath delay and angles of arrival of each of these resolvable multipath components. The angle of arrival estimation is achieved without using a large antenna array at the receiver, but with an iterative angle estimation mechanism that can work with just two receive antennas. The resulting range sensor can detect obstacles at multiple directions and multiple distances by analyzing just one received Wi-Fi packet. The generated obstacle map is then improved by combining different map instances estimated by the robot from different positions. Advantages of this sensor includes very fast scan time and, since we can reuse the ubiquitous Wi-Fi radio to transmit the OFDM signal on air, the sensor cost is much lower compared to traditional range sensors.