Vehicle Sensors Research Articles

Development of test methods for electromechanical sensors of vehicles used for operation in regions with a hot climate

Currently, there are various methods of testing vehicles and its components used by different car manufacturing companies. Also, these methods are given in the UNECE regulatory documents in the field of transport, but these documents do not take into account the separate operating conditions of cars in different regions. In this regard, the paper presents the road and climatic conditions of Central Asia characterized by peculiar characteristics. Based on the climatic features of the operation in a hot region, technical requirements have been developed for electromechanical sensors of cars for operation in hot climatic conditions, which should ensure the reliability of these components when the car runs at least 250,000 km. The method of testing electromechanical sensors of cars has been presented. It is established that in order to ensure the normal operability of sensors operating in hot climates, it is necessary to conduct tests at a nominal ambient temperature 45 ± 5 °C. The test procedure is built based on this parameter. All developed test methods are consistent with current standards in the field of electromagnetic, electromechanical and other electrical parameters of the automotive industry. Schematic diagrams of stands for testing have been given. The paper reflects the methodology of accelerated reliability tests, in which the temperature is set as the accelerating factor. The tests fully reflect the operational characteristics of the movement of cars in a sharply continental climate with a hot slope of ambient temperature with sharp temperature changes. After each test, it is determined that the nominal parameters of the sensors are checked for compliance with the established requirements. Conclusions reflecting the results of the work carried out have been made.

Smartphone-based hard-braking event detection at scale for road safety services

Road crashes are the sixth leading cause of lost disability-adjusted life-years (Dalys) worldwide. One major challenge in traffic safety research is the sparsity of crashes, which makes it difficult to achieve a fine-grain understanding of crash causations and predict future crash risk in a timely manner. Hard-braking events have been widely used as a safety surrogate due to their relatively high prevalence and ease of detection with embedded vehicle sensors. As an alternative to using sensors fixed in vehicles, this paper presents a scalable approach for detecting hard-braking events using the kinematics data collected from smartphone sensors. We train a Transformer-based machine learning model for hard-braking event detection using concurrent sensor readings from smartphones and vehicle sensors from drivers who connect their phone to the vehicle while navigating in Google Maps. The detection model shows superior performance with a 0.83 Area under the Precision–Recall Curve (Pr-auc), which is 3.8×better than a Gps speed-based heuristic model, and 166.6×better than an accelerometer-based heuristic model. The detected hard-braking events are strongly correlated with crashes from publicly available datasets, supporting their use as a safety surrogate. In addition, we conduct model fairness and selection bias evaluation to ensure that the safety benefits are equally shared. The developed methodology can benefit many safety applications such as identifying safety hot spots at road network level, evaluating the safety of new user interfaces, as well as using routing to improve traffic safety.

Cross-Area Travel Time Uncertainty Estimation From Trajectory Data: A Federated Learning Approach

Along with urbanization and the deployment of GPS sensors in vehicles and mobile phones, massive amounts of trajectory data have been generated for city areas. The analysis of these data has substantially contributed to research and advancements of travel time estimation. However, existing work focuses on estimating travel time inside a particular area, and cross-area travel time estimation has privacy security challenges due to data exchange issues among areas. Meanwhile, the majority of methods estimate a deterministic travel time for a given trajectory, which does not account for complex traffic situations and user requirements. To address these problems, we propose a cross-area travel time uncertainty estimation algorithm for estimating the uncertainty of travel times while preserving privacy among different areas. Specifically, we design a comprehensive cross-area privacy-preserving solution that trains a tailor-made neural network travel time estimator in each area by local data, and incorporates federated learning for training. Furthermore, we employ Bayesian deep learning principles and adopt Monte-Carlo dropout to quantify the uncertainty associated with travel time. To evaluate the proposed approach, we conduct a series of comprehensive case studies with two real-world trajectory datasets. Extensive results demonstrate the superiority of the proposed approach compared to baselines in the context of the cross-area setting.

A Novel Real-Time LiDAR Data Streaming Framework

Data transmission of familiar sensor data sources like video and audio is well explored. However, data acquired from the sensor emitting voluminous data cannot easily be transmitted due to a lack of bandwidth, which causes bottlenecks in the transmission. One valuable sensor in autonomous vehicles is light detection and ranging (LiDAR). In this article, a novel framework has been proposed, which is capable of efficiently streaming LiDAR data over the Internet. The proposed framework consists of various subsystems that handle data in multiple stages. On the sender side, the stages include data acquisition from the sensor, compression, encryption, and encoding of data, followed by transmission of encoded data over the Internet. On the receiver side, decoding, decryption, decompression, and data visualization are performed. The end-to-end framework runs over the robot operating system (ROS) platform. The data is acquired via ROS topics on the sender side and again published on ROS topics from the browser on the receiver side. This allows the users to apply any processing algorithm online with tolerable latency. The challenge of Clock Synchronization was addressed during the testing process. A detailed analysis to validate the quality of decompressed data has also been performed for which cloud distance was used as a metric. The implemented model was able to perform data transmission with an average latency of 160.62 ms for low-resolution compression mode.

Survey on Fish-Eye Cameras and Their Applications in Intelligent Vehicles

Fish-eye cameras have become essential sensors in intelligent vehicles. Due to its unique projection principle, a fish-eye camera can provide a large field of view. Benefiting from this special feature, fish-eye cameras have rich applications in intelligent vehicles. However, dataset and distortion problems are still challenges when applying fish-eye cameras in reality. This work introduces the projection principle of fish-eye cameras, and four classic fish-eye image representation models are presented. Then, the typical fish-eye datasets are presented, including real collected data and virtually generated data. Through the organization and summarization of the relevant studies, we demonstrate various applications of fish-eye cameras in intelligent vehicles, e.g., object detection and tracking, image segmentation, mapping and localization, and around-view monitoring. These works design various strategies to exploit the advantages of fish-eye cameras and prevent image distortion problems, showing the broad application prospects of such cameras. Finally, we discuss the development tendencies of intelligent vehicle applications involving fish-eye cameras.

Research on autonomous driving technology for a robot vehicle in mountainous farmland using the Quasi-Zenith Satellite System

The centimeter-level augmentation service (CLAS) provided by Michibiki (also known as the Quasi-Zenith Satellite System (QZSS)) can be utilized as a navigation sensor for a robot vehicle in orchards, which are often located in mountainous areas with steep slopes. Buildings, windbreaks, and the inclination of the antenna affected the CLAS less than they did the virtual reference station real-time kinematic (VRS-RTK) method, which is one of the most common navigation sensors for autonomous driving in agriculture. The time to first fix (TTFF) of the CLAS was 60 s earlier near buildings and 90 s earlier in case of a tilted antenna compared to that of the VRS-RTK. Moreover, the CLAS does not depend on an internet signal such as that provided for mobile broadband, since it uses augmentation information from the QZSS. A positioning test in a real vineyard, where grapevines are trained over vertical trellises, revealed that the CLAS could acquire FIX solutions constantly, whereas the VRS-RTK could not because of weak internet signals. The accuracy of the CLAS was less than 6.4 cm for a 95% confidence interval at static positioning and less than 3 cm root mean square (RMS) at dynamic positioning. A robot vehicle navigated by the CLAS autonomously traveled two paths, including a turning maneuver, with less than 9 cm RMS and less than 20 cm MAX. The results showed that the CLAS enabled the robot vehicle to work autonomously without any collisions in an actual vineyard.

A preliminary feasibility study of potential market applications for non-commercial technology magnets

Climate change has spurred increased electrification through means of transportation, hydropower, and wind turbines which has caused an increased demand for permanent magnet materials. Current commercial magnet technologies rely heavily on several critical materials such as neodymium, praseodymium, dysprosium, samarium, and cobalt which are primarily sourced and refined outside of the United States (U.S.). To combat these problems, the Critical Materials Institute (CMI) has begun research into alternative magnet compositions to reduce critical material content. Additionally, these alternative magnets can fulfill a gap in the market between high performance neodymium-iron-boron (Nd-Fe-B) and samarium cobalt (Sm–Co) magnets and low performance ferrite or bonded Nd-Fe-B magnets, earning the term gap magnets. This research seeks to compile a simple strategy for identifying an application for these alternative magnets and assessing preliminary market impacts through substitution for two example magnets. The first magnet was identified to be applicable for ancillary motors and sensors in conventional gasoline vehicles with a maximum substitution of 4,825 metric tonnes (mt) per year by the year 2050. The second magnet was identified to be applicable for magnetic couplings in energy and industrial sectors with a maximum substitution of 978 mt per year by the year 2050.

Real-Time Measurement of NOx Emissions from Modern Diesel Vehicles Using On-Board Sensors

In this work, the performance of on-board vehicle exhaust emission sensors is investigated and compared to reference laboratory and on-road instrumentation for two modern diesel light-duty commercial vehicles, type-approved as Euro 6d-TEMP-EVAP-ISC and Euro 6d-ISC-FCM. The first step of the analysis was to perform emissions tests in the laboratory and compare the NOx concentrations registered by the vehicle sensors available at the engine-out and tailpipe positions with those recorded by reference laboratory instrumentation. In a second step, tests were also conducted on road, comparing the performance of on-board sensors with those of Portable Emission Measurement System (PEMS) analysers, which were taken as references. The uncertainty related to exhaust flow measurements was also addressed. In particular, emissions factors calculated using the flow rates measured either in the laboratory or on-road were compared to those obtained by computing exhaust flows with on-board recorded data available from the vehicle electronic control unit. Results showed maximum deviations on the order of 34% in laboratory tests and of 21% during on-road measurements. Finally, measurements were also carried out during a diesel particulate filter regeneration event, showing the good performance of the on-board sensors even when high NOx concentrations were present. These conditions can be similar to those experienced in the case of an after-treatment system malfunction or of a high-emitting event, and can thus be of interest for real-time malfunction identification and monitoring.

Vehicle assisted bridge damage assessment using probabilistic deep learning

Vehicle assisted monitoring has shown promising potential for the condition assessment of existing bridges in a road network, by removing practical complications faced in traditional Structural health monitoring (SHM) methods such as traffic interruption and dense deployment of sensors. However, the combination of different measurement sources during vehicle assisted monitoring has not yet been fully explored. This paper aims to evaluate the potential benefit of considering multiple measured responses from various sources, including fixed sensors on the bridge and on-board vehicle sensors. To this end, this paper proposes a Probabilistic Deep Neural Network, a stochastic data-driven framework for damage assessment. This framework enables the combination of vehicle and bridge responses to extract damage sensitive features for the classification of different damage states. In addition, the proposed method estimates the uncertainty of its predictions, providing an indication of the reliability of the result. The proposed method is validated using two numerical based case studies while considering realistic operational conditions, which include temperature oscillations, additional traffic, and measurement noise. The results from this study indicate that combining multiple sensor information results in lower uncertainties in damage detection and localisation. The results also suggest that the proposed method is robust in handling measurement noise and varying environmental conditions.

Monitoring Air Quality in Urban Areas Using a Vehicle Sensor Network (VSN) Crowdsensing Paradigm

We present new advances in monitoring particulate matter (PM) in urban areas within a participatory vehicle sensor network (VSN) that exploits the use of multiple mobile low-cost IoT devices. These devices send geolocated PM measurements to an IT infrastructure and enabled us to reconstruct, in real time, the spatial and temporal distribution of pollutants in the study area in a web-based environment. The newly acquired data were integrated with independent reference measurements available from governmental environmental agencies. We deployed the infrastructure in the city of Trieste (Italy), since the beginning of 2021, with the help of several volunteers and the local transportation authority (Trieste Trasporti). By analysing the data, we delineate areas with lower air quality and identify the possible causes of these anomalies. We were able to define a belt outside the urban center where an enhanced concentration of pollutants occurs due to a higher flux of vehicular traffic that tends to jam there. Overall, our results demonstrate that this approach can be helpful in supporting urban planning and can also stimulate the community to reflect on how they can improve air quality in the area they live by reducing the use of private cars in favour of more widespread public transportation usage.

Deep reinforcement learning for adaptive path planning and control of an autonomous underwater vehicle

Research into intelligent motion planning methods has been driven by the growing autonomy of autonomous underwater vehicles (AUV) in complex unknown environments. Deep reinforcement learning (DRL) algorithms with actor-critic structures are optimal adaptive solutions that render online solutions for completely unknown systems. The present study proposes an adaptive motion planning and obstacle avoidance technique based on deep reinforcement learning for an AUV. The research employs a twin-delayed deep deterministic policy algorithm, which is suitable for Markov processes with continuous actions. Environmental observations are the vehicle's sensor navigation information. Motion planning is carried out without having any knowledge of the environment. A comprehensive reward function has been developed for control purposes. The proposed system is robust to the disturbances caused by ocean currents. The simulation results show that the motion planning system can precisely guide an AUV with six-degrees-of-freedom dynamics towards the target. In addition, the intelligent agent has appropriate generalization power.

A New Method for Analyzing the Aero-Optical Effects of Hypersonic Vehicles Based on a Microscopic Mechanism

Aero-optical effects are the key factors that restrict the accuracy of the optical sensors of hypersonic vehicles, and the numerical simulation of aero-optical effects is a powerful tool with which to analyze aero-optical distortion. Most existing research focuses on the simulation analysis of refraction distortion based on the density field at the macro level via the ray-tracing method. In this paper, a method for analyzing aero-optical effects based on the interaction between photons and gas molecules is proposed and can explain the optical distortion and energy dissipation caused by aero-optical effects at the micro level. By establishing a transmission model of photons in turbulence, a simulation method of aero-optical effects based on a microscopic mechanism is designed and breaks through the limitations of a traditional macro method in energy analyses. The optical distortion parameters based on photonics are compared with the physical quantities of traditional aero-optical effects, which verifies the effectiveness of the micro analysis on the macro scale and provides a new idea for studying the microscopic mechanism of aero-optical effects.

An in-tire-pressure monitoring SoC using FBAR resonator-based ZigBee transceiver and deep learning models

In automobiles, tires play an important role in ensuring safe driving. Hence tire pressure monitoring system (TPMS) has become the most important safety system in automobile vehicles. In this work, a new indirect TMPS is proposed that measures the pressure of all the tyres indirectly using the available sensors already mounted for typical vehicle dynamics control purposes. The available sensors in vehicles usually contain a power supply module, a microcontroller, and transceiver and resonator circuitry. The available wheel speed sensor is modified by reducing the wireless receiver system in communication protocols to minimize the power consumption. The transceiver structure is modified using a suitable local oscillator (LO) frequency plan based on a temperature-compensated thin film bulk acoustic wave resonator (FBAR). In addition, the proposed TPMS SoC uses a deep-stacked autoencoder to detect the adherence conditions of the tires and a pressure loss based on wheel speed. This module will initially pre-process the speed signals to remove the speed error caused by manufacturing error. After pre-processing, different features such as statistical, frequency-domain and the fitted frequency-domain characteristics are extracted from the wheel speed signal and fed to the learning model for the judgment of the tire state. The proposed system was designed in Verilog synthesis and layout were made in cadence tool. For ASIC implementation, we use TSMC 65 nm CMOS Library for LNA/hybrid mixer design. The performance measures of the pressure detection model in terms of accuracy and errors are evaluated. Also, the performance metrics of the transceiver architecture in terms of gain, NF, IIP3, S11 and FOM are evaluated. The experimental evaluation describes that the model procures a maximum gain of 58 dB, NF of 2.15 dB and IIP3 of 6.6 dBm respectively. The die area is 0.45 mm2with 0.8 supply voltage and 0.83 mW power. In addition to this, the relationship between peak amplitude and speed vehicle regarding time and frequency domain features are evaluated.

Smart Black Box DVR System in IT-Based Vehicle Emergency Rescue Environment

In this paper, the driving habits and patterns are analyzed by automatically recognizing vehicle accidents and recording event data when dangerous situations occur, applying connected car service through IT technology. In addition, SBDS designs the highly scalable modeling that reproduces videos which are recorded before and after accidents. SBDS provides emergency rescue and traffic information services by receiving IoT sensor data connected to have an embedded system in the vehicle and notifying the accident location and video information to the control center, using wireless network service even when the vehicle is dormant. In addition, it is possible to implement a wireless e-Call service by interlocking ECU information and vehicle sensor data in case an emergency occurs.

Intelligent Driving Vehicle Object Detection Based on Improved AVOD Algorithm for the Fusion of LiDAR and Visual Information

The detection of obstacles of intelligent driving vehicles becomes a primary condition to ensure safe driving. LiDAR and camera are the main sensors for intelligent vehicles to obtain information about their surroundings, and they each have their own benefits in terms of object detection. LiDAR can obtain the position and geometric structure of the object, and camera is very suitable for object recognition, but the reliance on environmental perception by a single-type sensor can no longer meet the detection requirements in complex traffic scenes. Therefore, this paper proposes an improved AVOD fusion algorithm for LiDAR and machine vision sensors. The traditional NMS (non-maximum suppression) algorithm is optimized using a Gaussian weighting method, while the 3D-IoU pose estimation loss function is introduced into the target frame screening module to upgrade the 2D loss function to 3D and design the 3D-IoU criterion. By comparing the detection accuracy of the algorithm proposed in this paper with that of the traditional method, it has been found that the improved AVOD fusion algorithm significantly improved the detection efficiency and the detection accuracy reached 96.3%. The algorithm proposed in this paper can provide a new approach for object detection of intelligent driving vehicles.

Measuring Roadway Lane Widths Using Connected Vehicle Sensor Data.

The United States has over three trillion vehicle miles of travel annually on over four million miles of public roadways, which require regular maintenance. To maintain and improve these facilities, agencies often temporarily close lanes, reconfigure lane geometry, or completely close the road depending on the scope of the construction project. Lane widths of less than 11 feet in construction zones can impact highway capacity and crash rates. Crash data can be used to identify locations where the road geometry could be improved. However, this is a manual process that does not scale well. This paper describes findings for using data from onboard sensors in production vehicles for measuring lane widths. Over 200 miles of roadway on US-52, US-41, and I-65 in Indiana were measured using vehicle sensor data and compared with mobile LiDAR point clouds as ground truth and had a root mean square error of approximately 0.24 feet. The novelty of these results is that vehicle sensors can identify when work zones use lane widths substantially narrower than the 11 foot standard at a network level and can be used to aid in the inspection and verification of construction specification conformity. This information would contribute to the construction inspection performed by agencies in a safer, more efficient way.

How to use cognitive tools to increase sustainability of elderly people’s mobility?

Aging of societies is a major international trend, thus effective and long-term development of activities for the elderly is an important issue. Vehicles must improve the range of activities of older people and increase their life trajectory beyond their age limits. With human participation, autonomous vehicles need to improve driving capabilities to drive safely in traffic scenarios and implement sustainable solutions. The discussion focuses on the impact on driving behavior, the functionality of vehicle sensors, and the interaction with traffic road users. This paper illustrates that autonomous driving tasks can benefit aging drivers due to vehicle sensors and systems, and road users when dealing with new or unexpected traffic situations. Identifying cognitive changes and relationships is important better to understand the road environment’s cognitive processes and behaviors.

Camera-LiDAR Fusion Method with Feature Switch Layer for Object Detection Networks.

Object detection is an important factor in the autonomous driving industry. Object detection for autonomous vehicles requires robust results, because various situations and environments must be considered. A sensor fusion method is used to implement robust object detection. A sensor fusion method using a network should effectively meld two features, otherwise, there is concern that the performance is substantially degraded. To effectively use sensors in autonomous vehicles, data analysis is required. We investigated papers in which the camera and LiDAR data change for effective fusion. We propose a feature switch layer for a sensor fusion network for object detection in cameras and LiDAR. Object detection performance was improved by designing a feature switch layer that can consider its environment during network feature fusion. The feature switch layer extracts and fuses features while considering the environment in which the sensor data changes less than during the learning network. We conducted an evaluation experiment using the Dense Dataset and confirmed that the proposed method improves the object detection performance.

Conceptualization, proposed design and theoretical investigations of an active adaptive cutting device with enhanced energy dissipation potential

Active adaptive energy dissipation systems can receive and interpret data from onboard vehicle sensors to respond to external stimuli by modifying their geometries and corresponding load bearing capacities. This is a highly promising subfield in structural crashworthiness since these devices possess multiple potential configurations, thus allowing them to achieve the preferred balance between high-capacity energy absorption and mitigation of human injury for a broad and disparate range of loading conditions. A novel cutting-based active adaptive energy absorber was conceptualized, design and theoretically assessed in this investigation with guidance from previous experimental observations and extensively validated modeling approaches. The proposed, hydraulically-driven active adaptive cutter (HAAC) was capable of transitioning between 4, 6, 8 and 12-bladed cutting deformation modes in either direction (e.g., from 6 to 12-bladed cutting or vice versa) and in a nondestructive manner. The steady-state force capacity could be varied by an average factor of 2.6, quantifying the significant range available to the HAAC regardless of extrusion material or geometry. A broad scale analytical study revealed that the proposed HAAC could eclipse the energy absorbing performance of the traditional axial crushing deformation mode for AA6061-T6 extrusions with 50.80 mm to 76.20 mm diameters and 1.59 mm to 3.18 mm wall thicknesses, with an average enhancement of 73 % calculated for the 12-bladed cutting mode. The newly proposed active adaptive capabilities allow the system to carefully balance energy absorption requirements and mitigation of occupant/pedestrian injury in a manner which the current state-of-the-art cannot achieve.

Distance to empty soft sensor for ford escape electric vehicle

Electric vehicle (EV) drivers require reliable distance to empty (DTE) indication to plan their trips. In the current study, feed forward neural networks based soft sensors were designed to accurately predict DTE in a Ford Escape EV. The proposed DTE soft sensors were trained on actual drive cycle data and rated DTE using Levenberg Marquardt, Bayesian Regularization and Scaled Conjugate Gradient algorithms. Regression models were also developed for comparisons. Primary results show that the Bayesian Regularization trained soft sensor network with eleven hidden layer neurons achieved the highest testing accuracy (99.64%) among the two layered networks, followed by the Levenberg Marquardt (two layered, eleven hidden layer neurons, testing accuracy 99.62%) and Scaled Conjugate Gradient trained networks (two layered, seven hidden layer neurons, testing accuracy 99.49%). The linear and non linear regression models attained 96.19% and 97.53% accuracies respectively. Deeper soft sensor networks yielded better prediction accuracies at higher computation times. The five layered Bayesian Regularization trained network (with ten neurons in each hidden layer) maximized DTE prediction accuracy to 99.89%, but at the cost of 1175% more training time as compared to the best performing two layered network soft sensor. An optimal choice of prediction accuracy considering reasonable computation timescales can help reduce range anxiety of EV users significantly.