Development of infrared reflectance characteristics of surrogate roadside objects

Abstract

An important topic in autonomous vehicle related research in recent times is road departure warning (RDW) and road keeping assistance (RKA). RDW or RKA should be able to recognize and avoid roadside objects. Standard tests are needed to evaluate the performance of RDW and RKA feature of cars from different manufacturers. To avoid damage to the cars under test and the test environment during testing, there is a need of soft, durable and reusable surrogate targets representing various real roadside objects such as curb, concrete divider and metal guardrail. These surrogate objects should have representative characteristics of real roadside objects from the point of view of various commonly used object detection sensors on the vehicles such as camera, radar and LIDAR. Transportation Active Safety Institute (TASI) at Indian University-Purdue University Indianapolis (IUPUI) is in the process of developing surrogate concrete divider, curb metal guardrail and grass that should be recognized as real roadside objects by LIDAR sensors, can be crashed without damage to the test vehicle and can be reused even after multiple crashes. The first step is to understand what the representative roadside objects should look like from the point of view of LIDAR units using laser of various wavelengths, and the next step is to design surrogate objects that successfully emulate the properties of the real roadside objects. Reflectance of an object is an important property for LIDAR detection. This thesis describes an approach for the determination of infrared reflectance property of concrete, metal guardrail and grass for different LIDAR view angles. Various samples of each of these roadside objects were evaluated. Based on these measurements, the suggested reflectance of surrogate roadside objects in the common LIDAR wavelength range of 800-1100 nm is specified. Finally, the design of surrogate roadside objects that satisfy these requirements is described, and the infrared reflectance of these surrogate objects are compared to the suggested reflectance bounds for different LIDAR view angles.