A Robot Localization Framework Using CNNs for Object Detection and Pose Estimation

TL;DR

This paper presents a CNN-based framework for robot localization and identification using external camera images, achieving high accuracy and real-time performance for various robot types and patterns.

Contribution

The authors introduce a novel two-stage CNN framework for robot detection, identification, and pose estimation, including a new training data generation process.

Findings

Achieved up to 98% mAP@IOU0.5 accuracy.

Orientation error as low as 1.6 degrees.

Operates at 50 Hz on a GPU.

Abstract

External localization is an essential part for the indoor operation of small or cost-efficient robots, as they are used, for example, in swarm robotics. We introduce a two-stage localization and instance identification framework for arbitrary robots based on convolutional neural networks. Object detection is performed on an external camera image of the operation zone providing robot bounding boxes for an identification and orientation estimation convolutional neural network. Additionally, we propose a process to generate the necessary training data. The framework was evaluated with 3 different robot types and various identification patterns. We have analyzed the main framework hyperparameters providing recommendations for the framework operation settings. We achieved up to 98% mAP@IOU0.5 and only 1.6{\deg} orientation error, running with a frame rate of 50 Hz on a GPU.