Non-Line-of-Sight Detection Based on Neuromorphic Time-of-Flight Sensing

Non-line-of-sight(NLOS) detection and ranging aim to identifyhidden objects by sensing indirect light reflections. Although numerouscomputational methods have been proposed for NLOS detection and imaging,the post-signal processing required by peripheral circuits remainscomplex. One possible solution for simplifying NLOS detection andranging involves the use of neuromorphic devices, such as memristors,which have intrinsic resistive-switching capabilities and can storespatiotemporal information. In this study, we employed the memristivespike-timing-dependent plasticity learning rule to program the time-of-flight(ToF) depth information directly into a memristor medium. By couplingthe transmitted signal from the source with the photocurrent fromthe target object into a single memristor unit, we were able to inducea tunable programming pulse based on the time interval between thetwo signals that were superimposed. Here, this neuromorphic ToF principleis employed to detect and range NLOS objects without requiring complexperipheral circuitry to process raw signals. We experimentally demonstratedthe effectiveness of the neuromorphic ToF principle by integratinga HfO2 memristor and an avalanche photodiode to detectNLOS objects in multiple directions. This technology has potentialapplications in various fields, such as automotive navigation, machinelearning, and biomedical engineering.