SLAM in Dynamic Environments: A Deep Learning Approach for Moving Object Tracking Using ML-RANSAC Algorithm.

Masoud S Bahraini,Mohammad Bozorg,Ahmad B Rad

doi:10.3390/s19173699

Abstract

The important problem of Simultaneous Localization and Mapping (SLAM) in dynamic environments is less studied than the counterpart problem in static settings. In this paper, we present a solution for the feature-based SLAM problem in dynamic environments. We propose an algorithm that integrates SLAM with multi-target tracking (SLAMMTT) using a robust feature-tracking algorithm for dynamic environments. A novel implementation of RANdomSAmple Consensus (RANSAC) method referred to as multilevel-RANSAC (ML-RANSAC) within the Extended Kalman Filter (EKF) framework is applied for multi-target tracking (MTT). We also apply machine learning to detect features from the input data and to distinguish moving from stationary objects. The data stream from LIDAR and vision sensors are fused in real-time to detect objects and depth information. A practical experiment is designed to verify the performance of the algorithm in a dynamic environment. The unique feature of this algorithm is its ability to maintain tracking of features even when the observations are intermittent whereby many reported algorithms fail in such situations. Experimental validation indicates that the algorithm is able to perform consistent estimates in a fast and robust manner suggesting its feasibility for real-time applications.

Highlights

Autonomous navigation in unknown and dynamic environments is a central problem in many robotic applications
We have presented a case whereby machine learning could be embedded within Simultaneous Localization and Mapping (SLAM) to address dynamic environments
We require an algorithm for perception and object detection to solve the SLAM problem

Summary

Introduction

Autonomous navigation in unknown and dynamic environments is a central problem in many robotic applications. If there are many moving objects in the environment, SLAM algorithms that are based on mapping static objects fail. The stationary as well as dynamic objects are predicted and updated using one estimator In such cases, we focus on obtaining a consistent map of static objects, discriminating between static and dynamic objects and concurrently estimate and track moving features. Using the ML-RANSAC algorithm, moving objects can be detected, while localization of the robot and mapping of stationary objects are in progress. In the case that data association can only be performed using the compatibility matrix, the RANSAC method are not used, which has usually a high level of computational efforts It results in executing the algorithm at a very high speed, in dynamic environments.

Related Works

ML-RANSAC

Results and Discussion

Pioneer

Conclusions