COMPUTER VISION BASED SYSTEM FOR AUTONOMOUS TAXIING OF AIRCRAFT

Prashant Gaikwat, Abhishek Mukhopadhyay, Anujith Muraleedharan, Mukund Mitra, Pradipta Biswas
Indian Institute of Science, Bengaluru
AVIATION Journal Vol 27 No 4 (2023)
Paper Code

Abstract

The system integrates both lane detection and collision detection and avoidance models. The lane detection component employs a segmentation model consisting of two parallel architectures. An airport dataset is proposed, and the collision detection model is evaluated with it to avoid collision with any ground vehicle. The lane detection model identifies the aircraft’s path and transmits control signals to the steer-control algorithm. The steer-control algorithm, in turn, utilizes a controller to guide the aircraft along the central line with 0.013 cm resolution. To determine the most effective controller, a comparative analysis is conducted, ultimately highlighting the Linear Quadratic Regulator (LQR) as the superior choice, boasting an average deviation of 0.26 cm from the central line. In parallel, the collision detection model is also compared with other state-of-the-art models on the same dataset and proved its superiority. A detailed study is conducted in different lighting conditions to prove the efficacy of the proposed system. It is observed that lane detection and collision avoidance modules achieve true positive rates of 92.59% and 85.19%, respectively.

Overview

Lane and object detection model

MY ALT TEXT

(a) Overall block diagram of the lane and object detection model. (b) Encoder-Decoder block. (c) Dilated Convolutional block. (d) Vanila architecture for predicting δ and 𝛾 as two weights to give weightage to different segmentation blocks.

Results

MY ALT TEXT

This work presents an end-to-end solution using state-of-the-art components for real-world aircraft navigation, contrasting with existing systems that operate in simulated environments. Traditionally, manual inspection is needed after landing before an aircraft follows the taxiway. Our approach integrates lane and object detection algorithms to automate this process, ensuring collision avoidance and precise stopping. The navigation algorithm controls steering and halts the aircraft if the taxiway ends or the lane is undetected.

Additional Results

MY ALT TEXT

The primary objective revolved around comparing the controllers' ability to effectively correct the Turtlebot's deviation from the central lane. To achieve this, we designed three distinct paths – left, right, and central – and systematically examined the performance of all four controllers across these paths. The findings indicated that the Linear Quadratic Regulator (LQR) exhibited the highest accuracy, with a mean error of 0.26 cm and a standard deviation of 0.94.

BibTeX

@article{Gaikwad2023,
        title={Developing a computer vision based system for autonomous taxiing of aircraft},
        author={Gaikwad, P. and Mukhopadhyay, A. and Muraleedharan, A. and Mitra, M. and Biswas, P.},
        journal={Aviation},
        volume={27},
        number={4},
        pages={248--258},
        year={2023},
        month={Dec},
        doi={10.3846/aviation.2023.20588}
        }