Autonomous driving requires efficient and safe decision making motion planning in dynamic uncertain environments. Future movement of surrounding vehicles is often difficult to represent. Besides, most existing studies consider planning/control separately. Both them may lead the oscillation unsafe for autonomous driving. This paper proposes an integrated framework with oscillation-free capability. The proposed approach overcomes shortcomings lane change/keeping maneuvers able to: i) make...

The tire-road peak adhesion coefficient (TRPAC) describes the tire limit that a road can provide. TRPAC is key parameter for precise vehicle motion control and an important basis decision-making planning of intelligent vehicles. Considering critical difficult problems in estimation TRPAC, such as slow convergence low accuracy, method based on fusion dynamics machine vision proposed this paper. Based observability theory nonlinear systems, local weak dynamics-based estimator analyzed to...

The tire-road peak adhesion coefficient (TRPAC) is defined as the ratio of to vertical load tire, which can characterize ability a tire adhere road. Reliable TRPAC estimation not only benefit vehicle active safety system, but also serve intelligent transportation system improve traffic participants. Considering problems low accuracy and poor real-time performance caused by low-quality sensor information in existing methods, fusion framework based on assessment multisource quality proposed...

Reinforcement Learning (RL) has shown excellent performance in solving decision-making and control problems of autonomous driving, which is increasingly applied diverse driving scenarios. However, a multi-attribute problem, leading to challenges achieving multi-objective compatibility for current RL methods, especially both policy execution iteration. On the one hand, common action space structure with single type limits flexibility or results large behavior fluctuations during execution....

This paper proposes an optimized trajectory planner and motion framework, which aim to deal with obstacle avoidance along a reference road for autonomous driving in unstructured environments. The planning problem is decomposed into lateral longitudinal sub-tasks the road. First, vehicle kinematic model coordinates established describe movement of vehicle. Then, nonlinear optimization based on space domain employed smooth Second, multilayered search algorithm applied lateral-space obstacles...

Abstract Vehicle mass is an important parameter for motion control of intelligent vehicles, but hard to directly measure using normal sensors. Therefore, accurate estimation vehicle becomes crucial. In this paper, a method based on fusion machine learning and dynamic model introduced. method, feedforward neural network (FFNN) used learn the relationship between other state parameters, namely longitudinal speed acceleration, driving or braking torque, wheel angular speed. dynamics-based...

Slip rate control is important in improving vehicle stability and driving efficiency. In this paper, a robust slip system designed for distributed drive electric vehicles that consists of two estimators multi-driving conditions, speed estimator, an anti-windup variable structure tracking controller. Because there no driven wheel four-in-wheel-motor vehicle, the small large rates are based on dynamic kinematic methods, respectively, which can switch according to conditions. The convergence...

Deep reinforcement learning (DRL) has become a powerful method for autonomous driving while often lacking safety guarantees. In this paper, we propose Risk-fused Constraint Reinforcement Learning (RCDRL) with D3QN network safe decision making in lane change maneuver. The problem is formulated as state-wise MDP (SCMDP), which embeds rule-based risk-fused module. We map the action to trajectory layer via polynomial curve-based planner, combined predicted trajectories of surrounding vehicles...

The tire-road peak adhesion coefficient (TRPAC), which cannot be directly measured by on-board sensors, is essential to road traffic safety. Reliable TRPAC estimation can not only serve the vehicle active safety system, but also benefit of other participants. In this paper, a fusion method considering model uncertainty proposed. Based on virtual sensing theory, an image-based estimator deep-learning and kinematic designed realize accurate classification surface condition will travel in...

The obstacle-avoidance problem of intelligent vehicles is one the challenges that we face in path planning. In order to tackle it, this paper proposes a real-time planning approach based on tentacle algorithm and B-spline curve. approach, firstly, some virtual tentacles are built represent precalculated paths ego vehicle at current speed. Secondly, it selects best among them, which required provide safe driving direction sampling area for generating paths. When drives along path, generated...

Based on active disturbance rejection control technique and characteristics of electric power steering, a steering angle tracking controller is designed, which consists an aligning moment estimator to deal with modeling error nonlinearity steering. The based extended state observer takes system friction differential drive torque, unique phenomenon in distributed vehicle, into consideration. According the estimated differentiator, designed nonlinear feedback feedforward compensation laws....

In order to avoid collisions in emergency conditions, the autonomous vehicle may be necessitated maneuvers up their handling limits. these scenarios, must able react quickly although longitudinal and lateral coupled nonlinear dynamics are significantly enhanced. To improve accuracy of trajectory tracking stability at limits, this paper presents a hierarchical motion control framework that prioritizing response. For control, nested architecture consists deviation yaw rate is designed...

In the automotive field, electro-hydraulic brake systems (EHB) has been developed to take place of vacuum booster, having advantage faster pressure built-up and continuously regulation. Pressure control, one most crucial issues be solved for EHB, is influenced by system nonlinearities (e.g., pressure-position, friction). Recently, there are a series studies that focus on this issue. However, control based estimation rarely investigated in previous literatures. order achieve cost-effective...

How to make a controller robust and stable reject the disturbance of uncertainty is an inevitable challenge. Aiming at addressing lateral control problem for autonomous road sweeper, heading-error-based first order linear active rejective (HFO-LADRC) proposed in this paper. To eliminate error heading same time, new model, called motion, Lyapunov function was employed explore convergence ability error. Since model order, ADRC designed as linear, each module HFO-LADRC has been devised detail....

An integrated multi-constraint nonlinear model predictive control method is proposed to solve the problem of active obstacle avoidance under high speed emergency conditions. The intervention and exit mechanism algorithm set through minimum braking distance information current state vehicle. According different scenarios, path tracking controller integrates into objective function by setting risk index function, so as realize multi-objective optimization calculate front steering angle....

This research explores the impact of pesticides on crop quality and food safety. The widespread use is a fundamental practice in modern agriculture, aimed at protecting crops from perils pests diseases. However, pesticide usage goes beyond pest control, affecting both Studies indicate that prolonged exposure to specific pesticides, such as organophosphates, can lead adverse changes aroma, taste, nutrient content crops, raising concerns about nutritional value pesticide-treated produce....

Anti-slip regulation (ASR) is one of the research focus in field active safety electric vehicle. An ASR algorithm adaptive to road condition proposed this paper based on 4WD vehicle with in-wheel motors. The controller anti-windup sliding mode control robust wheel parameter uncertainty. longitudinal velocity estimator fusion dynamics method and kinematics adopted reduce estimation error. slope estimated using recursive least square forgetting factor acceleration sensor information calibrated...