Vision-based reinforcement learning for lane-tracking control

<p class="Abstract">The present study focused on vision-based end-to-end reinforcement learning in relation to<strong> </strong>vehicle control problems such as lane following and collision avoidance. The controller policy presented in this paper is able to control a small-scale robot to follow the right-hand lane of a real two-lane road, although its training has only been carried out in a simulation. This model, realised by a simple, convolutional network, relies on images of a forward-facing monocular camera and generates continuous actions that directly control the vehicle. To train this policy, proximal policy optimization was used, and to achieve the generalisation capability required for real performance, domain randomisation was used. A thorough analysis of the trained policy was conducted by measuring multiple performance metrics and comparing these to baselines that rely on other methods. To assess the quality of the simulation-to-reality transfer learning process and the performance of the controller in the real world, simple metrics were measured on a real track and compared with results from a matching simulation. Further analysis was carried out by visualising salient object maps.</p>

Autonomous Driving Implementation in an Experimental Environment

Autonomous systems require identifying the environment and it has a long way to go before putting it safely into practice. In autonomous driving systems, the detection of obstacles and traffic lights are of importance as well as lane tracking. In this study, an autonomous driving system is developed and tested in the experimental environment designed for this purpose. In this system, a model vehicle having a camera is used to trace the lanes and avoid obstacles to experimentally study autonomous driving behavior. Convolutional Neural Network models were trained for Lane tracking. For the vehicle to avoid obstacles, corner detection, optical flow, focus of expansion, time to collision, balance calculation, and decision mechanism were created, respectively.

Mobile robot navigation based on line landmarks using the Braitenberg controller and image processing

Line following and lane tracking are robotic navigation techniques that use lines as a guide. The techniques can be applied to mobile robots in the industry. This research applied the Braitenberg controller and image processing to control and obtain line information around the mobile robot. The robot was implemented using Arduino Uno as a controller. A webcam was connected to a computer that performs image processing using canny edge detection and sends the data to the robot controller via serial communication. The robot can navigate on the side of the line, and the success rate of the system is 100 % at a turn of 135 ° and 80 % at a turn of 90 °.

Lane Tracking pada Robot Beroda Holonomic menggunakan Pengolahan Citra

ABSTRAK Robot merupakan teknologi yang dapat diterapkan bidang pertanian dan industri. Salah satu teknik navigasi robot yang dapat diterapkan di bidang pertanian dan industri adalah lane tracking. Untuk bernavigasi robot membutuhkan sebuah alat untuk mengenali lingkungannya, alat tersebut dapat berupa sensor atau kamera. Salah satu kelebihan menggunakan kamera jika dibandingkan dengan sensor adalah dapat mengurangi penggunaan perangkat keras untuk mengenali lingkungan robot. Fokus utama penelitian ini adalah membuat robot beroda holonomic untuk bernavigasi di antara dua garis yang berada di sebelah kiri dan kanan robot menggunakan kamera. Kamera digunakan untuk menangkap citra di depan robot, citra tersebut diolah disebuah SBC (Single Board Computer) untuk mendapatkan parameter jumlah pixel antara garis tengah robot dengan garis sebelah kanan dan kiri robot. Parameter tersebut kemudian diolah untuk menentukan kecepatan motor pada roda robot holonomic. Hasil yang diperoleh adalah dari setiap pengujian robot mampu bernavigasi pada jalur yang telah ditentukan. Kata kunci: Lane Tracking, Pengolahan Citra, Robot Beroda Holonomic ABSTRACT Robotic navigation techniques that can be applied in agriculture and industry is lane tracking. To navigate, robots need device to recognize the environment, the device can use sensors or cameras. The main focus of this research is to make holonomic wheeled robot to navigate between two lines located on the left and right of the robot using the camera. The camera is used to capture the image in front of the robot, the image is processed in an SBC (Single Board Computer) to get the paramters of the number of pixels between the center line with the right and left lines of the robot. These paramaters are the processed to determine the motor speed on the holonomic robot wheel. The result of each test is that the robot is able to navigate on a predetermined path. Keywords: Lane Tracking, Image Processing, Holonomic Wheeled Robot