An Embedded System Design to Build Real-Time 2D Maps for Unknown Indoor Environments

Abstract

This paper presents a remotely controllable, differentially driven wheeled mobile robot development inorder to build 2D maps for unknown indoor environments. This system would eliminate the need to preexploresuch environments. Main aim of the study is to develop a system with high accuracy by usingminimum number of sensors and a processor with low cost especially for comparatively small indoorareas. The distance traveled was calculated using the wheel odometry method. Obstacles surroundingthe robot, the distance traveled, and the robot’s orientation were obtained using an ultrasonic distancesensor, optical encoder, and a 3D orientation sensor (also known as an Attitude and Heading ReferenceSystem –AHRS), respectively. In addition, the characteristics of the system hardware components wereempirically explored, and the errors resulting from the sensors were evaluated. The non-linearitypercentage error arising from the encoder was defined and then compensated for. The hysteresisbehavior of the ultrasonic distance sensors was also empirically tested. All of the tasks were conductedby using a low-cost FPGA (Field Programmable Gate Arrays) board. A graphical development platformof National Instruments (NI) LabVIEW and its FPGA Module was preferred in the study for embeddedsystem programming instead of the text-based HDLs (Hardware Description Languages). Thisdistinguishes the proposed system from similar prior studies.