Predicting the Noise Covariance With a Multitask Learning Model for Kalman Filter-Based GNSS/INS Integrated Navigation

201

The Department of Transport in the United Kingdom recorded 25,080 motor vehicle fatalities in 2019. This situation stresses the need for an intelligent transport system (ITS) that improves road safety and security by avoiding human errors with the use of autonomous vehicles (AVs). Therefore, this survey discusses the current development of two main components of an ITS: (1) gathering of AVs surrounding data using sensors; and (2) enabling vehicular communication technologies. First, the paper discusses various sensors and their role in AVs. Then, various communication technologies for AVs to facilitate vehicle to everything (V2X) communication are discussed. Based on the transmission range, these technologies are grouped into three main categories: long-range, medium-range and short-range. The short-range group presents the development of Bluetooth, ZigBee and ultra-wide band communication for AVs. The medium-range examines the properties of dedicated short-range communications (DSRC). Finally, the long-range group presents the cellular-vehicle to everything (C-V2X) and 5G-new radio (5G-NR). An important characteristic which differentiates each category and its suitable application is latency. This research presents a comprehensive study of AV technologies and identifies the main advantages, disadvantages, and challenges.

Section: Calculation Of the Update Stagementioning

confidence: 99%

A Survey of Autonomous Vehicles: Enabling Communication Technologies and Challenges

Ahangar

Ahmed

Khan

et al. 2021

201

“…Another approach that uses reinforcement learning to adaptively estimate the process-noise covariance matrix was proposed by Gao et al [ 20 ], in which their algorithm used the deep deterministic policy gradient (DDPG) to extract the optimal process-noise covariance matrix estimation from the continuous action space, using an integrated navigation system as the environment and the reverse of the current positioning error as the reward. Wu et al [ 21 ] also proposed a deep learning framework combining a denoising autoencoder and a multitask temporal CNN. Multitask learning was used to optimize the loss for both the process-noise covariance and measurement-noise covariance matrices from KF simultaneously.…”

Section: Related Workmentioning

confidence: 99%

Trainable Quaternion Extended Kalman Filter with Multi-Head Attention for Dead Reckoning in Autonomous Ground Vehicles

Milam

Xie²,

Liu³

et al. 2022

Extended Kalman filter (EKF) is one of the most widely used Bayesian estimation methods in the optimal control area. Recent works on mobile robot control and transportation systems have applied various EKF methods, especially for localization. However, it is difficult to obtain adequate and reliable process-noise and measurement-noise models due to the complex and dynamic surrounding environments and sensor uncertainty. Generally, the default noise values of the sensors are provided by the manufacturer, but the values may frequently change depending on the environment. Thus, this paper mainly focuses on designing a highly accurate trainable EKF-based localization framework using inertial measurement units (IMUs) for the autonomous ground vehicle (AGV) with dead reckoning, with the goal of fusing it with a laser imaging, detection, and ranging (LiDAR) sensor-based simultaneous localization and mapping (SLAM) estimation for enhancing the performance. Convolution neural networks (CNNs), backward propagation algorithms, and gradient descent methods are implemented in the system to optimize the parameters in our framework. Furthermore, we develop a unique cost function for training the models to improve EKF accuracy. The proposed work is general and applicable to diverse IMU-aided robot localization models.

“…Theorem 1. (Bayesian Orthogonality Principle) If the weight function Ω Θ k,l of a linear filter satisfies (12), then the xα k obtained in (11) is called the optimal Bayesian least-squares estimation if and only if the following is the case.…”

Section: Kfpns Frameworkmentioning

confidence: 99%

“…As a result, the problem of designing a robust Kalman filter in practical applications where the knowledge of noise distribution is missing or imprecise is a big challenge for both researchers and developers. The authors of [7][8][9][10][11][12] proposed methods, which include so-called adaptive Kalman filtering, to estimate signal states and noise simultaneously, which works well when there is a lot of data used to obtain certain accurate performances in the entire estimated period. Later on, minimax-based Kalman filters [13][14][15] and finite impulse response Kalman filters [16][17][18][19] were presented.…”

Section: Introductionmentioning

confidence: 99%

An Overdispersed Black-Box Variational Bayesian–Kalman Filter with Inaccurate Noise Second-Order Statistics

Lin

Zhang

Zhao

et al. 2021

Aimed at the problems in which the performance of filters derived from a hypothetical model will decline or the cost of time of the filters derived from a posterior model will increase when prior knowledge and second-order statistics of noise are uncertain, a new filter is proposed. In this paper, a Bayesian robust Kalman filter based on posterior noise statistics (KFPNS) is derived, and the recursive equations of this filter are very similar to that of the classical algorithm. Note that the posterior noise distributions are approximated by overdispersed black-box variational inference (O-BBVI). More precisely, we introduce an overdispersed distribution to push more probability density to the tails of variational distribution and incorporated the idea of importance sampling into two strategies of control variates and Rao–Blackwellization in order to reduce the variance of estimators. As a result, the convergence process will speed up. From the simulations, we can observe that the proposed filter has good performance for the model with uncertain noise. Moreover, we verify the proposed algorithm by using a practical multiple-input multiple-output (MIMO) radar system.