Autonomous Parking Research Articles

An adaptive motion planning algorithm for obstacle avoidance in autonomous vehicle parking

In the recent era, machine learning-based autonomous vehicle parking and obstacle avoidance navigation have drawn increased attention. An intelligent design is needed to solve the autonomous vehicles related problems. Presently, autonomous parking systems follow path planning techniques that generally do not possess a quality and a skill of natural adapting behavior of a human. Most of these designs are built on pre-defined and fixed criteria. It needs to be adaptive with respect to the vehicle dynamics. A novel adaptive motion planning algorithm is proposed in this paper that incorporates obstacle avoidance capability into a standalone parking controller that is kept adaptive to vehicle dimensions to provide human-like intelligence for parking problems. This model utilizes fuzzy membership thresholds concerning vehicle dimensions and vehicle localization to enhance the vehicle’s trajectory during parking when taking into consideration obstacles. It is generalized for all segments of cars, and simulation results prove the proposed algorithm’s effectiveness.

Analysis and Improvement of Geometric Parallel Parking Methods with Respect to the Minimum Final Lateral Distance to the Parking Spot

Nowadays, the autonomous parking and park assist systems are getting more attention since they provide comfort and relief to the drivers and prevent parking-lot accidents. In this paper, we overview several main geometric path planning methods for autonomous parking of non-holonomic vehicles and we analyze their performances with respect to the final lateral distance to the parking spot. To the best of our knowledge, this kind analysis has never been done before for autonomous parallel parking algorithms. Moreover, we provide a new geometric method to reduce the final lateral distance between the vehicle and the parking spot. While the reference methods are insufficient for the final lateral distance, the proposed method introduces a realizable solution for long vehicles and enhances the performance for short vehicles. Simulations which are performed with various types of vehicles provide up to 48% better performance and show the efficiency of the new approach clearly.

Integration of Dimension-adaptive Obstacle Avoidance Algorithm in Fuzzy-based Autonomous Vehicle Parking

Sensor-based obstacle avoidance and Autonomous vehicle parking have been immensely researched in recent times. An integration of both will increase the usability of autonomous parking systems in dynamic and uncertain environments. The fuzzy logic theory is widely used to learn expert human skills for machines. However, existing fuzzy-based expert systems generally fail to mimic the natural adaptive skills of humans. The expert driver has a natural tendency to adapt to machine dynamics, especially vehicle-related. This paper proposes a novel non-holonomic dimension-based obstacle avoidance parking algorithm that integrates obstacle avoidance capabilities to a standalone parking controller. This algorithm is developed based on adaptive fuzzy membership inferences concerning passenger cars' different sizes and segments. It is tested for various vehicles in simulation results to show the effectiveness of the algorithm.

Visual SLAM in Long-Range Autonomous Parking Application Based on Instance-Aware Semantic Segmentation via Multi-Task Network Cascades and Metric Learning Scheme

<div class="section abstract"><div class="htmlview paragraph">Long-range Autonomous Parking is becoming an attractive application in terms of demands. The vehicle is capable of driving autonomously into the appointed parking slot when the driver leaves it at the drop-off spot. In this application, the ability of accurate localization has become a key issue, especially in GPS-denied environments. This paper proposes a method of localization and mapping for Long-range Autonomous Parking, which is achieved by Visual SLAM based on deep learning algorithms. Firstly, we propose an instance segmentation via multi-task network cascades, and even in a complex visual environment, the main roadway instances of interest in the parking lot IPM image can be detected, such as parking corners, speed bumps. Then we combine the information of wheel encoders to build a global semantic map of the parking lot. Vehicles can often rely on semantic map matching to achieve high-precision localization. However, without a good initial position, it is difficult to infer an accurate position by matching the semantic map, such as randomly selecting entrances to enter the parking lot. Therefore, we propose an area feature network based on metric learning to extract features that distinguish different areas and infer the approximate initial position of the vehicle. Specifically, we extract features from the images of the surround-view cameras, use the vehicle position as weak supervision, and finally construct an area feature map. In summary, our proposed method provides accurate vehicle localization and parking lot maps for Long-range Autonomous Parking.</div></div>

A Topological Map-based Path Coordination and Optimization Strategy for Autonomous Parking

A Topological Map-based Path Coordination and Optimization Strategy for Autonomous Parking

Autonomous Parking Trajectory Planning With Tiny Passages: A Combination of Multistage Hybrid A-Star Algorithm and Numerical Optimal Control

This paper introduces an autonomous parking trajectory planning method in an unstructured environment with narrow passages. The proposed hierarchical trajectory planner consists of a graph search layer and a numerical optimal control layer. The contribution mainly lies in the graph search layer, wherein a multistage hybrid A* algorithm is proposed to handle narrow passages formed by obstacles in the cluttered environment. In the multistage hybrid A* algorithm, a 2-dim A* search is conducted to find a global route that connects the starting and goal points. Along the derived global route, subtle segments that traverse narrow passages are extracted. Thereafter, the hybrid A* algorithm is used to plan kinematically feasible subpaths that connect the boundary points of each subtle segment. The hybrid A* algorithm is also used to find linking paths that connect adjacent subpaths. Combining all the subpaths and linking paths in a sequence yields a coarse path, which is converted into a coarse trajectory by attaching a time-optimal velocity profile to it. The coarse trajectory is fed into the numerical optimization layer as the initial guess. Simulation results indicate that the hierarchical trajectory planner runs much faster than prevalent ones in dealing with unstructured environments with narrow passages.

Proposal of Model Predictive Trajectory Planning Method for Autonomous Parking Considering Obstacle Avoidance Constraint with Coordinate Transformation

Proposal of Model Predictive Trajectory Planning Method for Autonomous Parking Considering Obstacle Avoidance Constraint with Coordinate Transformation

A guaranteed collision‐free trajectory planning method for autonomous parking

Planning a feasible, and safe trajectory is a crucial procedure for autonomous parking which remains to be fully solved. Generally, a trajectory is depicted by a sequence of configurations which include positions, and orientations. However, existing approaches usually design parking trajectories without considering the evolution from one configuration to the next. Hence, these trajectories are practically infeasible in some scenarios, and safety risks may exist. In this paper, a guaranteed collision-free trajectory planning method is proposed for autonomous parking. A dynamic optimisation problem is built, which includes vehicle kinematics, collision avoidance constraints, and physical restraints. Then the dynamic optimisation problem is discretised into a finite-dimensional nonlinear programming problem. To describe collision avoidance constraints, the concepts of the virtual protection frame, and the magnification parameter are introduced. It is highlighted that a novel collision detection criterion is raised from the view of geometry. Tests in three common scenarios, and a complex scenario illustrate the effectiveness of the proposed approach. Moreover, the stability of the algorithm is discussed in this paper. Via the proposed approach, an integrated design of planning, and control is possible.

Human‐like motion planning for autonomous parking based on revised bidirectional rapidly‐exploring random tree* with Reeds‐Shepp curve

AbstractAutonomous parking is one significant autonomous application and will be implemented in daily life in the near future. Due to encountered narrow environments, the issues related to autonomous parking, such as path quality requirements, strict collision avoidance, and motion direction changes, must be overcome properly. Moreover, to be applied in daily driving activities, real‐time planning and human preference should be fulfilled by the designed motion planners. Therefore, an efficient and human‐like motion planning method based on the revised Bidirectional Rapidly‐Exploring Random Tree* (Bi‐RRT*) with Reeds‐Shepp curve is presented. The proposed method results in human‐like paths which have high trajectory quality and consistency for parking scenarios due to the revised Bi‐RRT* framework. Strict collision checking model guarantees the resulting paths to be collision‐free and even leaves safe distance from obstacles and uncertainties. State space adjustment makes the path optimization more efficient and effective. On the other hand, the cost function revision makes the resulting paths meet human driving behavior, such as less backward driving and motion direction changes. In addition, rigorous simulations and analysis demonstrate the effectiveness of the cost function revision and the state space adjustment and illustrate good performance of the proposed approach in common and even complex parking scenarios.

Iterative-learning error compensation for autonomous parking of mobile manipulator in harsh industrial environment

In wide-area and multi-sites manufacturing scenarios, the mobile manipulator suffers from inadequate autonomous parking performance due to the harsh industrial environment. Instead of struggling to model various errors or calibrate multiple sensors, this paper resolves the above challenge by proposing an iterative-learning error compensation scheme that consists of offline pre-regulation and online compensation, which can improve the compensation efficiency and accommodate the error fluctuations caused by environmental fluctuations. Integrating an improved Monte-Carlo localization and eye-in-hand vision technique, an effective measurement system is firstly developed to accurately obtain the parking data without requiring superfluous facilities or cumbersome measurement. Then, after removing the data outliers utilizing the Grubbs test, offline pre-regulation is achieved to give a suitable initial value and increase the compensation convergence. To reduce the time-varying systematic errors and parking error fluctuations, online compensation is presented by offering an efficacious estimation of environmental fluctuations using fuzzy logic rules and providing an adaptive iterative-learning law. Finally, the feasibility and effectiveness of the presented compensation method are validated by extensive experiments implemented on a self-developed mobile manipulator.

Privacy-Aware Autonomous Valet Parking: Towards Experience Driven Approach

Driverless parking, an influential application of Mobility as a Service (MaaS) model, is one of the clear early benefits for autonomous vehicles, given often narrow spaces and multiple potential hazards (such as pedestrians stepping out from in between other vehicles). In recent years, real momentum has been building up for designing automated parking models for vehicles. However, in such an autonomous parking design, location privacy and identity privacy issues are always overlapping due to the improper sharing of data. Most existing studies barely investigate and poorly address such privacy issues. Motivated by this, we develop (and evaluate) an experience-driven, secure and privacy-aware framework of parking reservations for automated cars. Our idea of using differential privacy with zero-knowledge proof provides both security and privacy guarantees to users. Furthermore, the performance of the developed model is enhanced by exploiting reinforcement learning approach such that the utility of the system and the parking reservation rate can be maximized. Extensive evaluation demonstrates the superiority of the proposed model.

Creating rule-based agents for artificial general intelligence using association rules mining

In this paper, our focus is on using a rule-based approach to develop agents with artificial general intelligence. In rule-based systems, developing effective rules is a huge challenge, and coding rules for agents requires a large amount of manual work. Association rules mining (ARM) can be used for discovering specific rules from data sets and determining relationships between data sets. In this paper, we introduce a modified ARM method and use it to discover rules that analyse the surrounding environment and determine movements for an agent-guided vehicle that has been designed to achieve autonomous parking. The rules are created by our ARM-based method from training data gained during manual training in customised parking scenarios. In this system, data are represented in terms of fuzzy symbolic elements. We have tested our system by simulation in a virtual environment to demonstrate the effectiveness of this new approach.

A Comparative Study of Parking Path Following Methods for Autonomous Parking System

A Comparative Study of Parking Path Following Methods for Autonomous Parking System

An Autonomous Parking System of Optimally Integrating Bidirectional Rapidly-Exploring Random Trees and Parking-Oriented Model Predictive Control

Autonomous parking techniques can be used to tackle the lacking problem of parking spaces. In this paper, a sampling-based motion planner consisting of optimizing bidirectional rapidly-exploring random trees* (Bi-RRT*) and parking-oriented model predictive control (MPC) is proposed to properly deal with various parking scenarios. The optimal Bi-RRT* approach aims to improve the common defects of traditional sampling-based motion planners, such as uncertainties of path quality and consistency, and exploring inefficiency in narrow spaces. For this reason, the proposed motion planner is able to overcome strict environments with obstacles and narrow spaces. The parking-oriented MPC is then designed for steering and speed controls simultaneously for accurately and smoothly tracking parking paths. Furthermore, the proposed controller is dedicated to work under the practical scenarios, such as vehicle considerations, real-time control, and signal delay. To verify the effects of the proposed autonomous parking system, extensive simulations and experiments are conducted in common and strict parking scenarios, such as perpendicular parking, parallel parking. The simulation results not only verify the effects of each technical element, but also show the capability to deal with the various parking scenarios. Furthermore, various on-car experiments sufficiently demonstrate that the proposed system can be actually implemented in everyday life.

Investigating the state of connected and autonomous vehicles: a literature Review

Connected and autonomous vehicles (CAVs) are expected to be deployed in the near future, however it is still not clear whether the technology will provide more benefits than it will present drawbacks. This study aims to collate the research already completed in this domain and highlight an area of knowledge which is currently lacking sufficient investigation. The review shows that there is a wide range of topics and sub-topics covered in this area. We focused mainly on three key sub-topics that had fewest studies: adaptive tolling systems, autonomous parking facilities and predicting adoption over time. Past work on adaptive tolling has predominantly modelled situations where the traffic condition is uniform, whereas the impact of dynamic traffic and unexpected events or incidents has not yet been considered. Studies on autonomous parking facilities need to focus more on closely simulating normal operating conditions, such as having multiple operational vehicles, multiple parking spaces and a mix of CAVs and traditional vehicles. Although the positives of CAVs have been the predominant focus of discussion, the technology does present a number of uncertainties that make the full picture of its implementation and effects on its environments incomplete. An assumption made in many studies is that the market penetration of CAVs is 100%, however this will not be the case for some period of time. Therefore, studies need to consider the rate at which CAVs are adopted, which can be influenced by a number of factors.

Context-Based Parking Slot Detection With a Realistic Dataset

The autonomous parking of vehicles requires the ability to accurately locate an available parking slot in the vicinity of a vehicle. Since parking slots have a variety of shapes and colors, may be occluded by obstacles, or look different due to surroundings such as lighting, accurately locating them can be a challenging task. In this paper, we propose a context-based parking slot detection method inspired by the process of a human driver finding a parking slot. Our method consists of two deep network modules: a parking context recognizer and parking slot detector . The parking context recognizer identifies the parking environment (type, angle, and availability of a parking slot), whereas the parking slot detector locates the exact position of a parking slot by multiple type-based fine-tuned detectors with rotated anchor boxes and a rotated non-maximal suppression. In addition, we release a realistic parking slot dataset, which comprises 22817 images of parking slots having various attributes and external conditions. We also propose a new evaluation metric for parking slot detection, reflecting whether a vehicle can be parked within the detected parking slot. Through comparison and ablation study in experiments, we demonstrate that our method outperformed the previous deep-learning-based methods, along with having a short operation time. The source codes and the dataset are available at https://github.com/dohoseok/context-based-parking-slot-detect/ .

A bilevel optimal motion planning (BOMP) model with application to autonomous parking

In this paper, we present a bilevel optimal motion planning (BOMP) model for autonomous parking. The BOMP model treats motion planning as an optimal control problem, in which the upper level is designed for vehicle nonlinear dynamics, and the lower level is for geometry collision-free constraints. The significant feature of the BOMP model is that the lower level is a linear programming problem that serves as a constraint for the upper-level problem. That is, an optimal control problem contains an embedded optimization problem as constraints. Traditional optimal control methods cannot solve the BOMP problem directly. Therefore, the modified approximate Karush–Kuhn–Tucker theory is applied to generate a general nonlinear optimal control problem. Then the pseudospectral optimal control method solves the converted problem. Particularly, the lower level is the J_2-function that acts as a distance function between convex polyhedron objects. Polyhedrons can approximate objects in higher precision than spheres or ellipsoids. As a result, a fast high-precision BOMP algorithm for autonomous parking concerning dynamical feasibility and collision-free property is proposed. Simulation results and experiment on Turtlebot3 validate the BOMP model, and demonstrate that the computation speed increases almost two orders of magnitude compared with the area criterion based collision avoidance method.

A unified optimal planner for autonomous parking vehicle

In order to reduce the controlling difficulty caused by trajectory meandering and improve the adaptability to parking into regular lots, a versatile optimal planner (OP) is proposed. Taking advantage of the low speed specificity of parking vehicle, the OP algorithm was modeled the planning problem as a convex optimization problem. Collision-free constraints were formalized into the shortest distance between convex sets by describing obstacles and autonomous vehicle as affine set. Since employing Lagrange dual function and combining KKT conditions, the collision-free constraints translated into convex functions. Taking the national standard into account, 5 kinds of regular parking scenario, which contain 0°, 30°, 45°, 60° and 90° parking lots, were designed to verify the OP algorithm. The results illustrate that it is benefit from the continuous and smooth trajectory generated by the OP method to track, keep vehicle’s stability and improve ride comfort, compared with A* and hybrid A* algorithms. Moreover, the OP method has strong generality since it can ensure the success rate no less than 82% when parking planning is carried out at the start node of 369 different locations. Both of evaluation criteria, as the pear error and RMSE in x direction, y axis and Euclidean distance d, and heading deviation θ, are stable and feasible in real tests, which illustrates that the OP planner can satisfy the requirements of regular parking scenarios.

ITDP-Robot: Design of An Intelligent Transport Dispatch Parking Robot

In this paper, a novel electric autonomous parking robot prototype was proposed, which aims to address the parking hassle caused by the imbalance between the vehicle ownership and the amount of the parking spaces. The mechanical structure was elaborately designed to allow the parking robot to adapt to vehicles with different wheelbases and tracks. The electrical structure was constructed with the aim of X-by-wire and distributed component-based control concept. To be capable of autonomous driving, the parking robot software system based on ROS was designed with the capability of environment perception, self-localization and path planning. Furthermore, a simulation environment based on Gazebo was built in order to simplify the development of the parking robot’s autonomous driving algorithms and validate those algorithms’ robustness. Though this parking robot is under the prototype stage, the dispatch strategy and the convenience for parking were also considered. Compared with the state-of-art parking robot, this parking robot is not only capable of working indoor parking lots but also the complex outdoor environments.

Optimal control problem of multi‐vehicle cooperative autonomous parking trajectory planning in a connected vehicle environment

The development of intelligent network and intelligent vehicle technology makes it possible for multi-vehicle cooperative driving. In this study, the multi-vehicle cooperative autonomous parking condition is addressed by establishing a vehicle kinematics model combined with dynamic constraints and endpoint and collision avoidance constraints. The autonomous parking trajectory planning problem is transformed into an optimal control problem. The shortest parking completion time is set as the optimal cost function, and the optimal control problem is discretised by using the Gauss pseudo-spectral method. Then, the non-linear programming problem is solved. Moreover, an initial guess generation strategy is proposed, and the iteration method is used to solve the problem by adding collision avoidance constraints to the collision situation of the vehicle to improve the convergence and robustness of the problem. The simulation model is built in MATLAB/Simulink to simulate multi-vehicle cooperative autonomous parking under different working conditions. The simulation results show that the proposed method can solve the multi-vehicle cooperative autonomous parking trajectory planning problem uniformly and effectively. Compared with the traditional pseudo-spectral method, the proposed method has a faster convergence speed.