Autonomous Ground Vehicles Research Articles

Fast intent prediction of multi-cyclists in 3D point cloud data using deep neural networks

Inferring the intended actions of road-sharing users with autonomous ground vehicles in particularly vulnerable ones like cyclists is considered one of the tough tasks facing the wide-spread deployment of autonomous ground vehicles. One of the main reasons for that is the scarcity of the available datasets for that task due to the difficulty in obtaining those datasets in real environments. In this work, we first propose a pipeline that can synthetically produce 3D LiDAR data of cyclists hand-signalling a set of intended actions that are commonly done in real environments. Given the synthetically-produced labelled 3D LiDAR data sequences, we trained a framework that can simultaneously detect, track and give predictions about the intended actions of multi-cyclists in the scene on time. The proposed framework was evaluated using both synthetic and real data from a physical 3D LiDAR sensor. Our proposed framework has scored competitive and robust results in both synthetic and real environments with 88% in F1 measure with higher frame per second rate (12.9 FPS) than the 3D LiDAR sensor frame rate (10 Hz).

Delivery of Healthcare Resources Using Autonomous Ground Vehicle Convoy Systems: An Overview.

Utilizing military convoys in humanitarian missions allows for increased overall performance of healthcare logistical operations. To properly gauge performance of autonomous ground convoy systems in military humanitarian operations, a proper framework for comparative performance metrics needs to be established. Past efforts in this domain have had heavy focus on narrow and specialized areas of convoy performance such as human factors, trust metrics, or string stability analysis. This article reviews available Army doctrine for manned convoy requirements toward healthcare missions and establishes a framework to compare performance of autonomous convoys, using metrics such as spacing error, separation distance, and string stability. After developing a framework of comparison for the convoy systems, this article compares the performance of two autonomous convoys with unique convoy control strategies to demonstrate the application and utility of the framework.

Navigation and motion control techniques for surface unmanned vehicle and autonomous ground vehicle: a review

An excellent navigation, guidance, and control (NGC) system had a high impact on trajectory tracking and the following scenarios. Both scenarios will include the heading, tangent, and velocity parameters in the computation. However, the control system design problem is not a new issue in the unmanned surface vehicle (USV) and autonomous ground vehivle (AGV) due to this constraint faced by many researchers since early these autonomy developments. Hence, this paper listed and emphasizing the techniques, including techniques implementation, strength, and the algorithm's constraints, a fusion of several techniques implemented for vehicle's stability, a turning ahead, and heading estimation. This paper concerns the similar algorithm used in the USV and AGV. Most of the selected techniques are basic algorithms and have been frequently implemented to control both vehicles' systems. Previous research shows pure pursuit guidance is the most popular technique in AGV to control the degree-of-freedom (DOF) velocity and the dynamic rate (sway, surge, and yaw). Simultaneously, the line of sight (LOS) controller is very compatible with controlling the movement of the USV. In conclusion, the technique's simulation test needs further research that will expose in the actual situation.

Ontology-based approach to data exchanges for robot navigation on construction sites

As the use of autonomous Unmanned Ground Vehicles (UGV) for automated data collection from construction projects increases, construction stakeholders have become aware of a problem with inter-disciplinary semantic data sharing and exchanges between construction and robotic. Cross-domain data translation requires detailed specifications especially when it comes to semantic data translation. Building Information Modeling (BIM) and Geographic Information System (GIS) are the two digital building technologies used to capture and store semantic information for indoor structures and outdoor environments respectively. In the absence of a standard format for data exchanges between the construction and robotic domains, the tools of both industries have yet to be integrated into a coherent deployment infrastructure. In other words, the semantics of BIM-GIS cannot be automatically integrated by the robotic platforms currently being used. To enable semantic data transfer across domains, semantic web technology has been widely used in multi-disciplinary areas for interoperability. This paves the way to smarter, quicker and more precise robot navigation on construction sites. This paper develops a semantic web ontology integrating robot navigation and data collection to convey the meanings from BIM-GIS to the robot. The proposed Building Information Robotic System (BIRS) provides construction data that are semantically transferred to the robotic platform and can be used by the robot navigation software stack on construction sites. To meet this objective, first, knowledge representation between construction and robotic domains is bridged. Then, a semantic database integrated with the Robot Operating System (ROS) is developed, which can communicate with the robot and the navigation system to provide the robot with semantic building data at each step of data collection. Finally, the BIRS proposed system is validated through four case studies.

Robotics-based vineyard water potential monitoring at high resolution

The purpose of this research is deploying a proximal sensing solution using non-invasive and cost-effective sensors onboard an Autonomous Ground Vehicle (AGV) as a feasible way for building high-resolution maps of water potential in vineyards. The final objective is offering growers a practical system to make decisions about water management, especially for arid climatic conditions. The monitoring AGV was entirely developed within this research context, and as a result, it is a machine specifically designed to endure off-road conditions and harsh environments. The autonomous vehicle served as a massive, non-invasive, and on-the-go data collector robotic platform. The sensors used for measuring the relevant field variables were two spectral reflectance sensors (SRS), an infrared radiometer, and an on-board weather sensor. The collected data were displayed on comprehensible grid maps using the Local Tangent Plane (LTP) coordinate system. The proposed model has a coefficient of determination R2 of 0.69, and results from combining six parameters: the canopy and air temperatures (as the temperature difference), the relative humidity, the altitude difference, the Normalized Difference Vegetation Index (NDVI), and the Photochemical Reflectance Index (PRI). The strongest relationships found in this study were between the temperature difference and PRI, with an R2 of 0.75, and the temperature difference with the leaf water potential with an R2 of 0.61. The practical use of these high-resolution maps includes irrigation scheduling and harvest zoning for sorting grape quality, with a further use as inputs to complex artificial intelligence algorithms considering larger areas or complementing airborne data. Future improvements to make the models more robust and versatile will entail considering additional variables, locations, or grapevine cultivars, and even other crops grown in vertical trellis systems.

Design of a high-payload Mecanum-wheel ground vehicle (MWGV)

With the rapid developments of Industry 4.0 and Smart Manufacturing, customized manufacturing has been becoming greatly needed. Meanwhile, the challenge of production automation has become more bigger, especially for the automation of moving, picking, placing and manipulating objects. Many researchers have begun to work on Autonomous Ground Vehicles (AGVs). Most AGVs were utilized to carry middle or small objects, as the high-payload AGVs were rarely developed. This paper focused on the design of a High-Payload Mecanum-Wheel Ground Vehicle (MWGV), which was 1.7 m wide and 2.04 m long. The weight of the vehicle was 740 kg and it was able to carry the payload as its own weight (i.e. around 7,300 N). The safety factor of the structural strength was greater than 1.66 and the safety factor of the axial design was at least 6.24. The vehicle was designed to carry 150-kg weight with a reach of 1.375 m without falling. The design of Mecanum wheels provided great flexibility on movement with small rotational radius. Mathematical descriptions about how Mecanum wheels were controlled was also introduced in this paper. Furthermore, the mechatronics and software integrations were demonstrated. The final experimental results showed the developed MWGV was able to perform the desired movement properly.

Popov-H∞ Robust Path-Tracking Control of Autonomous Ground Vehicles with Consideration of Sector Bounded Kinematic Nonlinearity

Abstract This paper proposes a new approach to cope with the kinematic nonlinearity in the H∞ vehicle path-tracking controller synthesis problem. The kinematic nonlinearity presented in the vehicle lateral error state is found to satisfy the sector-bound condition. By isolating the sector bounded nonlinearity via an upper linear fractional transformation (LFT), a Lur'e system is formulated. A nominal robust controller is synthesized to meet both the Popov-H∞ criterion and the regional pole placement requirement. A polytopic gain-scheduling technique is subsequently employed to accommodate the effect of the varying vehicle longitudinal velocity. Finally, an instant-turning maneuver and a sharp lane-changing maneuver are tested in CarSim-Simulink joint simulations whose results demonstrate the superiority of the proposed Popov-H∞ controller over a conventional H∞ controller.

An Ellipsoidal Set-Membership Approach to Distributed Joint State and Sensor Fault Estimation of Autonomous Ground Vehicles

This paper is concerned with the problem of distributed joint state and sensor fault estimation for autonomous ground vehicles subject to unknown-but-bounded (UBB) external disturbance and measurement noise. In order to improve the estimation reliability and performance in cases of poor data collection and potential communication interruption, a multi-sensor network configuration is presented to cooperatively measure the vehicular yaw rate, and further compute local state and fault estimates. Toward this aim, an augmented descriptor vehicle model is first established, where the unknown sensor fault is modeled as an auxiliary state of the system model. Then, a new distributed ellipsoidal set-membership estimation approach is developed so as to construct an optimized bounding ellipsoidal set which guarantees to contain the vehicle's true state and the sensor fault at each time step despite the existence of UBB disturbance and measurement noises. Furthermore, a convex optimization algorithm is put forward such that the gain matrix of each distributed estimator can be recursively obtained. Finally, simulation results are provided to validate the effectiveness of the proposed approach.

Robotic Exploration of an Unknown Nuclear Environment Using Radiation Informed Autonomous Navigation

This paper describes a novel autonomous ground vehicle that is designed for exploring unknown environments which contain sources of ionising radiation, such as might be found in a nuclear disaster site or a legacy nuclear facility. While exploring the environment, it is important that the robot avoids radiation hot spots to minimise breakdowns. Broken down robots present a real problem: they not only cause the mission to fail but they can block access routes for future missions. Until now, such robots have had no autonomous gamma radiation avoidance capabilities. New software algorithms are presented that allow radiation measurements to be converted into a format in which they can be integrated into the robot’s navigation system so that it can actively avoid receiving a high radiation dose during a mission. An unmanned ground vehicle was fitted with a gamma radiation detector and an autonomous navigation package that included the new radiation avoidance software. The full system was evaluated experimentally in a complex semi-structured environment that contained two radiation sources. In the experiment, the robot successfully identified both sources and avoided areas that were found to have high levels of radiation while navigating between user defined waypoints. This advancement in the state-of-the-art has the potential to deliver real benefit to the nuclear industry, in terms of both increased chance of mission success and reduction of the reliance on human operatives to perform tasks in dangerous radiation environments.

Consistent Right-Invariant Fixed-Lag Smoother with Application to Visual Inertial SLAM

State estimation problems without absolute position measurements routinely arise in navigation of unmanned aerial vehicles, autonomous ground vehicles, etc., whose proper operation relies on accurate state estimates and reliable covariances. Unaware of absolute positions, these problems have immanent unobservable directions. Traditional causal estimators, however, usually gain spurious information on the unobservable directions, leading to over-confident covariance inconsistent with actual estimator errors. The consistency problem of fixed-lag smoothers (FLSs) has only been attacked by the first estimate Jacobian (FEJ) technique because of the complexity to analyze their observability property. But the FEJ has several drawbacks hampering its wide adoption. To ensure the consistency of a FLS, this paper introduces the right invariant error formulation into the FLS framework. To our knowledge, we are the first to analyze the observability of a FLS with the right invariant error. Our main contributions are twofold. As the first novelty, to bypass the complexity of analysis with the classic observability matrix, we show that observability analysis of FLSs can be done equivalently on the linearized system. Second, we prove that the inconsistency issue in the traditional FLS can be elegantly solved by the right invariant error formulation without artificially correcting Jacobians. By applying the proposed FLS to the monocular visual inertial simultaneous localization and mapping (SLAM) problem, we confirm that the method consistently estimates covariance similarly to a batch smoother in simulation and that our method achieved comparable accuracy as traditional FLSs on real data.

Observer-based finite frequency [formula omitted] state-feedback control for autonomous ground vehicles

This paper studies the problem of path tracking of Autonomous Ground Vehicles (AGVs) in the presence of sideslip angles. An observer is designed to estimate both the sideslip angle and the vehicle yaw rate, based on which an observer-based controller is established such that the closed-loop system is stable and the vehicle follows a desired path accurately. In particular, the nonlinear vehicle dynamics model is reformulated as a Linear Parameter Varying (LPV) system, and a finite frequency H∞ criteria is satisfied such that the disturbances are attenuated effectively, the parameter-dependent gain matrices are calculated simultaneously by solving a convex optimization problem. Simulation results show the effectiveness of the method proposed.

Model Predictive Path Planning for an Autonomous Ground Vehicle in Rough Terrain

An autonomous ground vehicle (AGV) in rough terrain typically experiences uncertain environment. Because the uncertainty makes overall performance of autonomous navigation degrade, the AGV requires a suitable path to maintain or improve the performance against the uncertainty. In order to handle this problem, this study proposes a model predictive path planning algorithm by employing a passivity-based model predictive control (MPC) optimization setup. The model predictive path planning method is formulated as a finite optimization problem with an objective function and several constraints. In the cost function, environment perception result about the AGV’s own neighborhood is included and the only traversable region has low cost value. To reflect dynamic characteristics of the AGV, the proposed method utilizes dynamic and kinematic models of the AGV as equality constraints and limited range of states and control input as inequality constraints. In addition, the stability of the path planning method is improved by a passivity constraint. The solution of the optimization problem is obtained using the particle swarm optimization (PSO) method. Finally, field tests are conducted to validate the performance of the proposed algorithm, and satisfactory results of the autonomous navigation were obtained.

Robust adaptive non-singular fast terminal sliding-mode lateral control for an uncertain ego vehicle at the lane-change maneuver subjected to abrupt change

This paper investigates a new approach to deal with the path-tracking for an uncertain autonomous ground vehicle (AGV) exposed to the unknown external disturbances caused by abrupt changes in the road bank angle. A robust adaptive non-singular fast terminal sliding mode control (RANFTSMC) is designed for the lane change maneuver (LCM) lateral control. Initially, a reference LCM path is planned by a recent procedure. Then, assuming that the upper bound of unknown external disturbances and uncertainties is known, a robust steering controller is implemented via non-singular fast terminal sliding mode control (NFTSMC) to ensure the vehicle lateral stability. However, the upper limit of the vehicle uncertainties and perturbations is unknown in real-world driving situations, therefore, an adaptive tuning law is developed to estimate this unknown upper bound. As long as the efficiency of the controller depends on their parameters, a meta-heuristic optimization algorithm is employed to deliver the controller optimal settings. The effectiveness of the proposed RANFTSMC is proven with theoretical illustrations and numerical simulation results. The offered steering controller maintains the lateral stability even in severe conditions where the vehicle longitudinal speed is up to 120 km/h. In this connection, the suggested strategy is compared with the higher-order sliding mode controller (HOSMC) and conventional sliding mode controller (SMC).

Spatial Path Tracking Controllers for Autonomous Ground Vehicles: Conventional and Nonconventional Schemes

Unlike time-based path tracking controllers, the [Formula: see text]-controller is a spatial path tracking controller. It is a purely geometric path tracking controller and essentially a P-controller to maintain the reasonable spatial distance, [Formula: see text], from the vehicle to the desired path. In this paper, we present some enhancement schemes using the non-conventional PI control laws via optimization. We propose to use a nonlinear term [Formula: see text] for the proportional controller. A fractional-order integral used to achieve a PI[Formula: see text] control. Among the schemes, an optimization search procedure applied to find optimal controller gains by meshing the regions around the values from approximate linear designs. The performance index for parametric optimization is the integration of the absolute purely spatial deviation from the desired path. Three different types of road shape were chosen and the Gazebo-ROS simulation results were presented to show the effectiveness of the proposed enhancement schemes. The results show that in some cases a smaller [Formula: see text] and [Formula: see text] can be achieved by using [Formula: see text] controller, but its disadvantage is there may be some oscillation. For PI[Formula: see text] controller, there is an additional adjustable parameter [Formula: see text], better performance can be achieved without significant disadvantages which is worth in-depth research.

Autonomous Ground Vehicle Lane-Keeping LPV Model-Based Control: Dual-Rate State Estimation and Comparison of Different Real-Time Control Strategies.

In this contribution, we suggest two proposals to achieve fast, real-time lane-keeping control for Autonomous Ground Vehicles (AGVs). The goal of lane-keeping is to orient and keep the vehicle within a given reference path using the front wheel steering angle as the control action for a specific longitudinal velocity. While nonlinear models can describe the lateral dynamics of the vehicle in an accurate manner, they might lead to difficulties when computing some control laws such as Model Predictive Control (MPC) in real time. Therefore, our first proposal is to use a Linear Parameter Varying (LPV) model to describe the AGV’s lateral dynamics, as a trade-off between computational complexity and model accuracy. Additionally, AGV sensors typically work at different measurement acquisition frequencies so that Kalman Filters (KFs) are usually needed for sensor fusion. Our second proposal is to use a Dual-Rate Extended Kalman Filter (DREFKF) to alleviate the cost of updating the internal state of the filter. To check the validity of our proposals, an LPV model-based control strategy is compared in simulations over a circuit path to another reduced computational complexity control strategy, the Inverse Kinematic Bicycle model (IKIBI), in the presence of process and measurement Gaussian noise. The LPV-MPC controller is shown to provide a more accurate lane-keeping behavior than an IKIBI control strategy. Finally, it is seen that Dual-Rate Extended Kalman Filters (DREKFs) constitute an interesting tool for providing fast vehicle state estimation in an AGV lane-keeping application.

Path Tracking Control of Autonomous Vehicles Subject to Deception Attacks via a Learning-Based Event-Triggered Mechanism.

This article investigates the problem of event-triggered secure path tracking control of autonomous ground vehicles (AGVs) under deception attacks. To relieve the burden of the shareable vehicle communication network and to improve the tracking performance in the presence of deception attacks, a learning-based event-triggered mechanism (ETM) is proposed. Different from existing ETMs, the triggering threshold of the proposed mechanism can be dynamically adjusted with conditions of the latest vehicle state. Each vehicle in this study is deemed as an agent, under which a novel control strategy is developed for these autonomous agents with deception attacks. With the assistance of Lyapunov stability theory, sufficient conditions are obtained to guarantee the stability and stabilization of the overall system. Finally, a simulation example is provided to demonstrate the effectiveness of the proposed theoretical results.

Game Theory-Based Control Strategy For Trajectory Following of Four-Wheel Independently Actuated Autonomous Vehicles

Trajectory Following Control (TFC) and lateral stability control are always significant for Autonomous Ground Vehicles (AGVs). However, sometimes it is quite a challenge to achieve excellent tracking performance with less stability loss. In this paper, a novel two-agent non-cooperative game framework is built to explore the coordination approach of the integrated TFC and Direct Yaw Control (DYC) issue. Then the game theory-based control strategies are proposed and two information patterns are investigated. In the open-loop case, the Nash equilibrium of each control cycle is determined by solving discrete-time Riccati equations. When it comes to the closed-loop pattern, the game complexity increases because each player can revise his own game strategy according to the other players’ strategy evolution, and thus the dynamic programming method based on the Bellman's principle is applied to derive the control strategies. Besides, a Linear Parameter-Varying (LPV) method is adopted to facilitate the process of discretizing the state-space model in every control cycle. Finally, simulation tests based on Carsim-Simulink joint platform and experimental tests are carried out to verify the effectiveness of the proposed approach.

Path planning for autonomous ground vehicles based on quintic trigonometric Bézier curve

Path planning is one of the essential steps for autonomous ground vehicles or even wheeled mobile robots. This paper proposes a general framework for the path planning using quintic trigonometric Bezier curve with two shape parameters and $$C_{3}$$ continuity. We express that when there are obstacles, predefined path can be adjusted by only using shape parameters without altering any obstacle. Additionally, the velocity, lateral acceleration, longitudinal and lateral jerks of the predefined cubic and quintic Bezier, and cubic and quintic trigonometric Bezier paths are compared. Also, a path surface for autonomous ground vehicles can be generated using developable surfaces.

Autonomous Ground Vehicle Path Planning in Urban Environments Using GNSS and Cellular Signals Reliability Maps: Models and Algorithms

In this article, autonomous ground vehicle (AGV) path planning is considered. The AGV is assumed to be equipped with receivers capable of producing pseudorange measurements to overhead global navigation satellite systems (GNSS) satellites and to cellular base stations in its environment. Parameters of the cellular pseudoranges related to the transmitter clock bias are estimated in an initialization step in an open-sky environment. The AGV fuses these pseudoranges to produce an estimate about its own states. The AGV is also equipped with a three-dimensional building map of the environment. Starting from a known starting point, the AGV desires to reach a known target point by taking the shortest distance, while minimizing the AGV's position estimation error and guaranteeing that the AGV's position estimation uncertainty is below a desired threshold. Toward this objective, a so-called signal reliability map is first generated, which provides information about regions where large errors due to poor GNSS line-of-sight or cellular signal multipath are expected. The vehicle uses the signal reliability map to calculate the position mean-squared error (MSE). An analytical expression for the AGV's state estimates is derived, which is used to find an upper bound on the position bias due to multipath. An optimal path planning generation approach, which is based on Dijkstra's algorithm, is developed to optimize the AGV's path while minimizing the path length and position MSE, subject to keeping the position estimation uncertainty and position estimation bias due to multipath below desired thresholds. The path planning approach yields the optimal path together with a list of feasible paths and reliable GNSS satellites and cellular base stations to use along these paths.

Autonomous Ground Vehicle Path Planning in Urban Environments Using GNSS and Cellular Signals Reliability Maps: Simulation and Experimental Results

A framework for autonomous ground vehicle (AGV) path planning using global navigation satellite systems (GNSS) signals and cellular long-term evolution (LTE) signals is evaluated through several simulations and experiments. The objective of path planning is to prescribe the optimal path for the AGV to follow to reach a desired target point. Optimality is defined as the shortest distance, while minimizing the AGV's position estimation error and guaranteeing that the uncertainty about its position is below a desired threshold. Path planning is prescribed via signal reliability maps, which provide information about regions where large errors due to cellular signal multipath or obstructed GNSS line-of-sight are expected. Simulation results are presented demonstrating that utilizing ambient cellular LTE signals together with GNSS signals 1) reduces the uncertainty about the AGV's position, 2) increases the number of feasible paths to choose from, which could be useful if other considerations arise (e.g., traffic jams and road blockages due to construction), and 3) yields significantly shorter feasible paths, which would otherwise be infeasible with GNSS signals alone. Experimental results on a ground vehicle navigating in downtown Riverside, CA, USA, are presented demonstrating a close match between the simulated and experimental results.