Search Results

The fusion of multi-modal perception in autonomous driving plays a pivotal role in vehicle behavior decision-making. However, much of the previous research has predominantly focused on the fusion of Lidar and cameras. Although Lidar offers an ample supply of point cloud data, its high cost and the substantial volume of point cloud data can lead to computational delays. Consequently, investigating perception fusion under the context of 4D millimeter-wave radar is of paramount importance for cost reduction and enhanced safety. Nevertheless, 4D millimeter-wave radar faces challenges including sparse point clouds, limited information content, and a lack of fusion strategies. In this paper, we introduce, for the first time, an approach that leverages Graph Neural Networks to assist in expressing features from 4D millimeter-wave radar point clouds. This approach effectively extracts unstructured point cloud features, addressing the loss of object detection due to sparsity.

The temperature monitoring of the lithium-ion battery is crucial for the advanced battery thermal management systems (BTMS) to improve performance and ensure operational safety and reliability of the battery system. In real applications, the core temperature of the battery is unfortunately unmeasurable due to the impracticality of placing a sensor inside the core, and has to be estimated online in real-time. Meanwhile, only the measurement of battery surface temperature can not meet the need for advanced BTMS due to the impact of the large temperature gradient between the surface and internal in high power applications. The battery core temperature estimation will become challenging when encountering sensor bias and noise.

To reduce the energy consumption level of electric vehicles, the working range of the regenerative braking system will gradually expand to the high state of charge of the battery. The time delay in the control signal transmission path of the high state of charge regenerative braking control process will affect the regenerative braking. At the same time, regenerative braking under a high state of charge puts forward higher requirements for the control accuracy of regenerative current. In the research of this paper, the motor model, battery model, and vehicle dynamics model are firstly established by using MATLAB/Simulink, and the dynamic relationship between regenerative current and regenerative braking torque is analyzed at the same time. Considering the system time delay, this paper proposes a high-charge regenerative braking control strategy (SPPC) that combines Smith prediction and prescribed performance control.

Highly autonomous vehicles have drawn the interests of many researchers in recent years. For highly autonomous vehicles, a high-definition (HD) map is crucial since it provides accurate information for autonomous driving. However, due to the possible fast-changing environment, the performance of HD maps will deteriorate over time if timely updates are not ensured. Therefore, this paper studies the updating of lightweight HD maps in closed areas. Firstly, a novel two-layer map model called a lightweight HD map is introduced to support autonomous driving in a flexible and efficient way. Secondly, typical updating of scenarios in closed areas with non-paved roads is abstracted into operations including area border expansion, road addition, and road deletion. Meanwhile, a map updating framework is proposed to address the issue of map updating in closed areas. Finally, an experiment is conducted to demonstrate the feasibility and effectiveness of the proposed map updating approach.

Panoramic images can provide critical information for Advanced Driving Assistance Systems (ADAS), such as parking spaces and surrounding vehicles. However, the vehicle in the bird's-eye view image is severely distorted and incomplete, and the visual information becomes very blurred in some illumination insufficient environments. If the driver cannot see the surrounding environment information, the risk of collision will increase, especially during parking. To better percept the local environment with the help of panoramic images, we use panoramic image segmentation results to construct a virtual surround view monitoring system to provide drivers with clearer perception information. Firstly, a lightweight segmentation network is redesigned based on SegNet, which will improve the accuracy of the segmentation without increasing the model’s inference time. Secondly, we build an augment visualization around view monitor (AV-AVM) system with regards to the segmentation results.

As one of the most challenging driving tasks, parking is a common but particularly troublesome problem in large cities. Recently, an excellent solution-automated valet parking (AVP) has become a hot research topic, which allows the driver to leave the vehicle in a drop-off area, while the vehicle driving into the parking slot by itself. For AVP, the precise localization is an indispensable module. However, the global positioning system (GPS) cannot be used in the underground parking lot and the localization method based on lidar is too expensive. In response to solve this problem, we propose a simultaneous localization and mapping system with the semantic information of parking slots (PS-SLAM), which is based on visual-inertial and around view images. First, the calibration of multi-sensors is conducted to obtain their intrinsic and extrinsic parameters. In this way, the around view image and transformation matrices between sensors can be acquired.

Due to the poor ability of millimeter wave radar in recognizing distant static objects, target loss and incomplete information will occur when it recognizes the static target in front, thus increasing the false alarm rate and missing alarm rate of the radar-dependent driving assistant system, which will reduce the driving safety and the acceptability of the assistant system. Aiming at the radar's poor ability to recognize static targets, this paper uses a model based on machine learning algorithm to recognize and track targets. The radar signals are collected and processed in different conditions, and the results show that the radar has a poor recognition effect when the distance is more than 100 meters and the speed is more than 19m/s.

As a new type of energy, free-piston linear generator (FPLG) attracts more research on its stable operation and power performance, while less on its combustion and emission performance. So, in this paper, the emission characteristics of FPLG in two different modes are studied through a port fuel injection (PFI) mode which was verified by the experiment and a gasoline direct injection (GDI) mode. The results showed that: both the GDI mode and the PFI mode produced large amounts of nitrogen oxide (NOx) during the working process. But the GDI mode produced before the PFI mode and it produced nearly 2 times than the PFI mode. However, the formation rate of NOx in GDI mode is much lower than that in PFI mode. Meanwhile, in both modes, 90% of NOX was generated in the cylinder at the temperature higher than 1750K, and only about 10% of NOX was generated at a temperature lower than 1750K.

The deformation of injector components cannot be disregarded as the pressure of the system increases. Deformation directly affects the characteristics of needle movement and injection quantity. In this study, structural deformation of the nozzle, the needle and the control plunger under different pressures is calculated by a simulation model. The value of the deformation of injector components is calculated and the maximum deformation location is also determined. Furthermore, the calculated results indicates that the deformation of the control plunger increases the control chamber volume and the cross-section area between the needle and the needle seat. A MATLAB model is established to The influence of structural deformation on needle movement characteristics and injection quantity is investigate by a numerical model. The results show that the characteristic points of needle movement are delayed and injection quantity increases due to the deformation.

The around view monitor (AVM) system for vehicles usually suffers from the distortion of surrounding objects caused by incomplete rectification and stitching, which seriously affects the driver's judgment of the surrounding environment during the reversing process. In response to solve this problem, an augmented around view monitor (AAVM) system fusing image and depth information is proposed, which highlights the point clouds of persons or vehicles at the rear of the vehicle. First, an around view image is generated from four fisheye cameras. Then, the calibration of multi TOF cameras is conducted to improve their accuracy of depth estimation and obtain extrinsic camera positions. Next, the 2D-driven object point cloud detection method is proposed to localize and segment object point clouds like vehicles or persons.

This paper intends to present a novel optimal trajectory planning method for obstacle avoidance on highways. Firstly, a mapping from the road Cartesian coordinate system to the road Frenet-based coordinate system is built, and the path lateral offset in the road Frenet-based coordinate system is represented by a function of quintic polynomial respecting the traveled distance along the road centerline. With different terminal conditions regarding its position, heading and curvature of the endpoint, and together with initial conditions of the starting point, the path planner generates a bunch of candidate paths via solving nonlinear equation sets numerically. A path selecting mechanism is further built which considers a normalized weighted sum of the path length, curvature, consistency with the previous path, as well as the road hazard risk.

Hydrogen fuel is a future energy to solve the problems of energy crisis and environmental pollution. Hydrogen internal combustion engines can combine the advantage of hydrogen without carbon pollution and the main basic structure of the traditional engines. However, the power of the port fuel injection hydrogen engines is smaller than the same volume gasoline engine because the hydrogen occupies the volume of the cylinder and reduces the air mass flow. The turbocharger can increase the power of hydrogen engines but also increase the NOx emission. Hence, a comprehensive controlling strategy to solve the contradiction of the power, BTE and NOx emission is important to improve the performance of hydrogen engines. This paper shows the controlling strategy for a four-stroke, 2.3L hydrogen engine with a turbocharger. The controlling strategy divides the operating conditions of the hydrogen engine into six parts according to the engine speeds and loads.

The driver's face detection and tracking method important for Advanced Driver Assistance Systems (ADAS) and autonomous driving in various situations. The deep SORT algorithm has integrated appearance information, the motion model and the intersection-over-union (IOU) distance methods, and has been applied to face tracking, but it depends on detection information in every frame. Once the detection information lacks, the deep SORT algorithm will wait until the target detects bounding boxes appear again, even if the target didn’t disappear or shield. Hence, we propose to use a new tracker that not completely depend on the detection algorithm to cascade with the deep SORT algorithm to realize stable driver's face tracking. At first, the driver's face detection and tracking will be accomplished by the MTCNN-deep-SORT algorithm.

Automated parking is an efficient way to solve parking difficulties and path planning is of great concern for parking maneuvers [1]. Meanwhile, the starting region of path planning greatly affects the parking process and efficiency. The present research of the starting region are mostly determined based on a single algorithm, which limits the flexibility and efficiency of planning feasible paths. This paper, taking parallel parking and vertical parking for example, proposes a method to calculate the starting region and select the most suitable path planning algorithm for parking, which can improve the parking efficiency and reduce the complexity. The collision situations of each path planning algorithm are analyzed under collision-free conditions based on parallel and vertical parking. The starting region for each algorithm can then be calculated under collision-free conditions.

HTL is a kind of biodiesel converted from wet biowaste via hydrothermal liquefaction (HTL), which has drawn increasing attention in recent years due to its wide range of raw materials (algae, swine manure, and food processing waste). However, from the previous experiments done in a constant volume chamber, it was observed that the presence of 20% of HTL in the blend produced as much soot as pure diesel at in chamber environment oxygen ratio of 21%, and even more soot at low oxygen ratios. It was also observed that n-butanol addition could reduce the soot emission of diesel significantly under all tested conditions. In this work, the spray and combustion characteristics of HTL and diesel blends with n-butanol added were investigated in a constant volume chamber. The in-chamber temperature and oxygen ranged from 800 to 1200 K and 21% to 13%, respectively, covering both conventional and low-temperature combustion (LTC) regimes.

Flash boiling has drawn much attention recently for its ability to enhance spray atomization and vaporization, while providing better fuel/air mixing for gasoline direct injection engines. However, the behaviors of flash boiling spray with multi-component fuels have not been fully discovered. In this study, isooctane, ethanol and the mixtures of the two with three blend ratios were chosen as the fuels. Measurements were performed with constant fuel temperature while ambient pressures were varied to adjust the superheated degree. Macroscopic and microscopic characteristics of flash boiling spray were investigated using Diffused Back-Illumination (DBI) imaging and Phase Doppler Anemometry (PDA). Comparisons between flash boiling sprays with single component and binary fuel mixtures were performed to study the effect of fuel properties on spray structure as well as atomization and vaporization processes.

The spray structure and vaporization processes of flash-boiling sprays in a constant volume chamber under a wide range of superheated conditions were experimentally investigated by a high speed imaging technique. The Engine Combustion Network’s Spray G injector was used. Four fuels including gasoline, ethanol, and gasoline-ethanol blends E30 and E50 were investigated. Spray penetration length and spray width were correlated to the degree of the superheated degree, which is the ratio of the ambient pressure to saturated vapor pressure (pa/ps). It is found that parameter pa/ps is critical in describing the spray transformation under flash-boiling conditions. Three distinct stages namely the slight flash-boiling, the transition flash-boiling, and the flare flash-boiling are identified to describe the transformation of spray structures.

There are a few different ways in which biofuels can be sourced, with the most popular coming from agricultural sources. An alternative approach is to utilize biowaste. An estimated 20 million dry tons of volatile organic compounds, or biowaste, is annually deposited in US municipal wastewaters. Most of this biowaste energy content is not recovered and, as a result, the biowaste could be a massive potential source of renewable energy. Biocrude diesel is converted from wet biowaste via hydrothermal liquefaction (HTL). Three types of feedstocks (algae, swine manure, and food processing waste) were converted into biocrude oil via HTL. From the previous experiments done in an AVL 5402 single-cylinder diesel engine, it was observed that the presence of 20% of HTL in the blend performed similarly during combustion to pure diesel. By studying these mixtures in a constant volume chamber, these observations could be compared to the results in the diesel engine.

The study of vehicle coupling state estimation accuracy especially in observer-based vehicle chassis control for improving road handling and ride comfort is a challenging task for vehicle industry under various driving conditions. Due to a large amount of life safety arising from vehicle roll behavior, how to precisely acquire vehicle roll state and rapidly provide for the vehicle control system are of great concern. Simultaneously, uncertainty is unavoidable for various aspects of a vehicle system, e.g., varying sprung mass, moment of inertia and position of the center of gravity. To deal with the above issues, a novel dual observer approach, which combines adaptive Unscented Kalman Filter (AUKF) and Takagi-Sugeno (T-S), is proposed in this paper. A full-car nonlinear model is first established to describe vehicle lateral and vertical coupling roll behavior under various road excitation.

The around view monitor (AVM) system for the long-body road vehicle with multiple articulated carriages usually suffers from the incomplete distortion rectification of fisheye cameras and the irregular image stitching area caused by the change of relative position of the cameras on different carriages while the vehicle is in motion. In response to these problems, a set of calibration and stitching methods of AVM are proposed. First, a radial-distortion-based rectification method is adopted and improved. This method establishes two lost functions and solves the model parameters with the two-step optimization method. Then, AVM system calibration is conducted, and the perspective transformation matrix is calculated. After that, a static basic look-up table is generated based on the distortion rectification model and perspective transformation matrix.