Object detection for KRSBI robot soccer using PeleeNet on omnidirectional camera

Abstract

Kontes Robot Sepak Bola Indonesia (KRSBI) is an annual event for contestants to compete their design and robot engineering in the field of robot soccer. Each contestant tries to win the match by scoring a goal toward the opponent's goal. In order to score a goal, the robot needs to find the ball, locate the goal, then kick the ball toward goal. We employed an omnidirectional vision camera as a visual sensor for a robot to perceive the object's information. We calibrated streaming images from the camera to remove the mirror distortion. Furthermore, we deployed PeleeNet as our deep learning model for object detection. We fine-tuned PeleeNet on our dataset generated from our image collection. Our experiment result showed PeleeNet had the potential for deep learning mobile platform in KRSBI as the object detection architecture. It had a perfect combination of memory efficiency, speed and accuracy.

A review on object detection for autonomous mobile robot

“…Instead of using multiple sensors, the study by [38] proposed an omnidirectional vision camera as a visual sensor for a robot to recognize the object's information. Furthermore, the proposed system utilizes PeleeNet as a deep learning model for object detection.…”

“…Object detection favours deep learning as it practices low-level feature development before critical enhancement. As computer vision involves extracting features from an image, known as image classification, the variety of prints allows it to adapt conveniently as inputs to deep learning [38], [42]. Nevertheless, the deep neural network architectures have varied based on performance detection due to the presence of different detectors, single-stage and two-stage detectors.…”

Bilateral control system of the space robot with large delays

“…Let the following initial and final values of the generalized coordinates of a free-flying space robot [22], [23] be given (Figure 7): Let us present a computer simulation of the forward kinematics problem for the FANUC ground manipulator [24], [25]. Take the following coordinates as the initial and final values: q 0 = 0, q 𝑘 = [3𝜋/2 −𝜋/8 𝜋/3 𝜋/6 −𝜋/2 𝜋/4] 𝑇 the dependence of the generalized coordinates on time will be similar.…”

Sistem Deteksi Bola pada Robot Kiper Pemain Sepakbola Beroda

“…Hal tersebut didukung dengan adanya kegiatan Kontes Robot Indonesia (KRI) kategori Kontes Robot Sepak Bola Indonesia (KRSBI) setiap tahunnya di Indonesia untuk menampung dan meningkatkan minat, bakat, dan kemampuan peneliti [5]. Kontes Robot Sepak Bola Indonesia (KRSBI) diadakan setiap tahun oleh Kementerian Riset, Teknologi, dan Pendidikan Tinggi Republik Indonesia (KEMENRISTEKDIKTI) mulai tahun 2017 [6]. Kontes Robot Sepak Bola Indonesia Beroda diselenggarakan menyesuaikan kondisi di Indonesia, misalnya untuk ukuran lapangan dan lain sebagainya [7].…”

“…Beberapa metode yang digunakan untuk pendeteksian bola sebagai objek yang dideteksi telah dilakukan sebelumnya, yaitu penggunaan Local Binary Pattern (LBP) [4], PeleeNet [6], dan Neural Network [15] yang mana masih membutuhkan perangkat keras yang mumpuni agar dapat mencapai performa yang diinginkan. Pada penelitian ini digunakan metode HoughCircle dan approxPolyDP [5] untuk mendeteksi bola, yang menjadi sebuah solusi yang dapat digunakan pada perangkat keras yang tidak memerlukan spesifikasi yang tinggi.…”