This paper aims to review the state of the art of Light Detection and Ranging (LiDAR) sensors for automotive applications, and particularly for automated vehicles, focusing on recent advances in the field of integrated LiDAR, and one of its key components: the Optical Phased Array (OPA). LiDAR is still a sensor that divides the automotive community, with several automotive companies investing in it, and some companies stating that LiDAR is a "useless appendix". However, currently there is not a single sensor technology able to robustly and completely support automated navigation. Therefore, LiDAR, with its capability to map in 3 dimensions (3D) the vehicle surroundings, is a strong candidate to support Automated Vehicles (AVs). This manuscript discusses the Society of Automotive Engineers (SAE) levels of autonomy for vehicles, highlighting current sensor challenges, and it analyses the strength and weaknesses of the sensor technologies currently deployed in vehicles to perceive the environment. Then, the manuscript discusses the main LiDAR technologies emerging in automotive, and focuses on integrated LiDAR, challenges associated with light beam steering on a chip, the use of Optical Phased Arrays, finally discussing current factors hindering the affirmation of silicon photonics OPAs and their future research directions.

The global Connected and Autonomous Mobility industry is growing at a rapid pace. To ensure the successful adoption of connected automated mobility solutions, their safety, reliability and hence the public acceptance are paramount. It is widely known that in order to demonstrate that L3+ automated systems are safer with respect to human drivers, upwards of several millions of miles need to be driven. The only way to efficiently achieve this amount of tests in a timely manner is by using simulations and high fidelity virtual environments. Two key components of being able to test an automated system in a synthetic environment are validated sensor models and noise models for each sensor technology. In fact, the sensors are the element feeding information into the system in order to enable it to safely plan the trajectory and navigate. In this paper, we propose an innovative real-time LiDAR sensor model based on beam propagation and a probabilistic rain model, taking into account raindrop distribution and size. The model can seamlessly run in real-time, synchronised with the visual rendering, in immersive driving simulators, such as the WMG 3xD simulator. The models are developed using Unreal engine, therefore demonstrating that gaming technology can be merged with the Automated Vehicles (AVs) simulation toolchain for the creation and visualization of high fidelity scenarios and for AV accurate testing. This work can be extended to add more sensors and more noise factors or cyberattacks in real-time simulations.

Assisted and automated driving functions in vehicles exploit sensor data to build situational awareness, however, the data amount required by these functions might exceed the bandwidth of current wired vehicle communication networks. Consequently, sensor data reduction, and automotive camera video compression need investigation. However, conventional video compression schemes, such as H.264 and H.265, have been mainly optimised for human vision. In this paper, we propose a semantic-aware (SA) video compression (SAC) framework that compresses separately and simultaneously region-of-interest and region-out-of-interest of automotive camera video frames, before transmitting them to processing unit(s), where the data are used for perception tasks, such as object detection, semantic segmentation, etc. Using our newly proposed technique, the region-of-interest (ROI), encapsulating most of the road stakeholders, retains higher quality using lower compression ratio. The experimental results show that under the same overall compression ratio, our proposed SAC scheme maintains a similar or better image quality, measured accordingly to traditional metrics and to our newly proposed semantic-aware metrics. The newly proposed metrics, namely SA-PSNR, SA-SSIM, and iIoU, give more emphasis to ROI quality, which has an immediate impact on the planning and decisions of assisted and automated driving functions. Using our SA-X264 compression, SA-PSNR and SA-SSIM have an increase of 2.864 and 0.008 respectively compared to traditional H.264, with higher ROI quality and the same compression ratio. Finally, a segmentation-based perception algorithm has been used to compare reconstructed frames, demonstrating a 2.7% mIOU improvement, when using the proposed SAC method versus traditional compression techniques.

Assisted and automated driving functions are increasingly deployed to support improved safety and efficiency and enhance the driver experience. However, there are still key technical challenges that need to be overcome, such as the degradation of perception sensor data due to noise factors. The quality of data being generated by sensors can directly impact the planning and control of the vehicle, which can affect the vehicle safety. This work builds on a recently proposed framework, analyzing noise factors on automotive light detection and ranging (LiDAR) sensors, and deploys it to camera sensors, focusing on the specific disturbed sensor outputs via a detailed analysis and classification of automotive camera-specific noise sources (30 noise factors are identified and classified in this work). Moreover, the noise factor analysis has identified two omnipresent and independent noise factors (i.e., obstruction and windshield distortion). These noise factors have been modeled to generate noisy camera data; their impact on the perception step, based on deep neural networks (NNs), has been evaluated when the noise factors are applied independently and simultaneously. It is demonstrated that the performance degradation from the combination of noise factors is not simply the accumulated performance degradation from each single factor, which raises the importance of including the simultaneous analysis of multiple noise factors. Thus, the framework can support and enhance the use of simulation for the development and testing of automated vehicles through careful consideration of the noise factors affecting camera data.

Automated vehicles (AVs) are one of the breakthroughs of this century. The main argument to support their development is increased safety and reduction of human and economic losses; however, to demonstrate that AVs are safer than human drivers billions of miles of testing are required. Thus, realistic simulation and virtual testing of AV systems and sensors are crucial to accelerate the technological readiness. In particular, perception sensor measurements are affected by uncertainties due to noise factors; these uncertainties need to be included in simulations. This work presents a framework to exhaustively analyze and simulate the effect of the combination of noise factors on sensor data. We applied the framework to analyze one sensor, the LiDAR (Light Detection and Ranging), but it can be easily adapted to study other sensors. Results demonstrate that single noise factor analysis gives an incomplete knowledge of measurement degradation and perception is dramatically hindered when more noises are combined. The proposed framework is a powerful tool to predict the degradation of AV sensor performance.

Situational awareness based on the data collected by the vehicle perception sensors (i.e. LiDAR, RADAR, camera and ultrasonic sensors) is key for achieving assisted and automated driving functions in a safe and reliable way. However, the data rates generated by the sensor suite are difficult to support over traditional wired data communication networks on the vehicle, hence there is an interest in techniques that reduce the amount of sensor data to be transmitted without losing key information or introducing unacceptable delays. These techniques must be analysed in combination with the consumer of the data, which will most likely be a machine learning algorithm based on deep neural networks (DNNs). In this paper we demonstrate that by compression tuning the DNNs (i.e. transfer learning by re-training with compressed data) the DNN average precision and recall can significantly improve when uncompressed and compressed data are transmitted. This improvement is achieved independently from the compression standard used for compression-training (i.e. AVC and HEVC), and also when training and transmitted data use the same compression standard or different compression standards. Furthermore, the performance of the DNNs is stable when transmitting data with increasing lossy compression rate, up to a compression ratio of approximately 160:1; above this value the performance starts to degrade. This work paves the way for the use of compressed sensor data in automated driving in combination with the optimisation of compression-tuned DNNs.