Poster: Lightweight Features Sharing for Real-Time Object Detection in Cooperative Driving

in 2023 IEEE Vehicular Networking Conference (VNC) (2023, April 26)

In model partitioning for real-time object detection, part of the model is deployed on a vehicle, and the remaining layers are processed in the cloud. Model partitioning requires transmitting intermediate ... [more ▼]

In model partitioning for real-time object detection, part of the model is deployed on a vehicle, and the remaining layers are processed in the cloud. Model partitioning requires transmitting intermediate features to the cloud, which can be problematic, given that the latency requirements are strict. This paper addresses this issue by demonstrating a lightweight featuresharing strategy while investigating a trade-off between detection quality and latency. We report details on layer partitioning, such as which layers to split in order to achieve the desired accuracy. [less ▲]

The Ugly Truth of Realistic Perception in Vehicular Simulations

Poster (2023, January 08)

Automated vehicles use sensors to perceive the environment, and studies have shown the limitations of these sensors. The onboard sensors may not detect objects when other participants occlude the Field of ... [more ▼]

Automated vehicles use sensors to perceive the environment, and studies have shown the limitations of these sensors. The onboard sensors may not detect objects when other participants occlude the Field of View (FoV). Thus, sensor efficiency must be tested to ensure its reliability. Simulation is an excellent test option due to the complexity associated with practical experiments. However, emerging simulation frameworks still have various limitations, especially when it comes to large-scale evaluation. This work investigates realistic perception simulation options for autonomous vehicles. We report the perception accuracy for different traffic scenarios. [less ▲]

FastCycle: A Message Sharing Framework for Modular Automated Driving Systems

E-print/Working paper (2022)

Automated Driving Systems (ADS) have rapidly evolved in recent years and their architecture becomes sophisticated. Ensuring robustness, reliability and safety of performance is particularly important. The ... [more ▼]

Automated Driving Systems (ADS) have rapidly evolved in recent years and their architecture becomes sophisticated. Ensuring robustness, reliability and safety of performance is particularly important. The main challenge in building an ADS is the ability to meet certain stringent performance requirements in terms of both making safe operational decisions and finishing processing in real-time. Middlewares play a crucial role to handle these requirements in ADS. The way middlewares share data between the different system components has a direct impact on the overall performance, particularly the latency overhead. To this end, this paper presents FastCycle as a lightweight multi-threaded zero-copy messaging broker to meet the requirements of a high fidelity ADS in terms of modularity, real-time performance and security. We discuss the architecture and the main features of the proposed framework. Evaluation of the proposed framework based on standard metrics in comparison with popular middlewares used in robotics and automated driving shows the improved performance of our framework. The implementation of FastCycle and the associated comparisons with other frameworks are open sourced. [less ▲]

Connected Vehicle Platforms for Dynamic Insurance

in in Proceedings of the 6th International Conference on Intelligent Traffic and Transportation (2022)

Refining Weakly-Supervised Free Space Estimation through Data Augmentation and Recursive Training

in Proceedings of BNAIC/BeneLearn 2021 (2021, November 12)

Free space estimation is an important problem for autonomous robot navigation. Traditional camera-based approaches rely on pixel-wise ground truth annotations to train a segmentation model. To cover the ... [more ▼]

Free space estimation is an important problem for autonomous robot navigation. Traditional camera-based approaches rely on pixel-wise ground truth annotations to train a segmentation model. To cover the wide variety of environments and lighting conditions encountered on roads, training supervised models requires large datasets. This makes the annotation cost prohibitively high. In this work, we propose a novel approach for obtaining free space estimates from images taken with a single road-facing camera. We rely on a technique that generates weak free space labels without any supervision, which are then used as ground truth to train a segmentation model for free space estimation. We study the impact of different data augmentation techniques on the performances of free space predictions, and propose to use a recursive training strategy. Our results are benchmarked using the Cityscapes dataset and improve over comparable published work across all evaluation metrics. Our best model reaches 83.64% IoU (+2.3%), 91:75% Precision (+2.4%) and 91.29% Recall (+0.4%). These results correspond to 88.8% of the IoU, 94.3% of the Precision and 93.1% of the Recall obtained by an equivalent fully-supervised baseline, while using no ground truth annotation. Our code and models are freely available online. [less ▲]

Towards a Framework to Evaluate Cooperative Perception for Connected Vehicles

in Proceedings of the 13th IEEE Vehicular Networking Conference 2021 (2021)

Over the past few years, Connected and Autonomous Vehicles (CAVs) have gained significant research attention. With the recent deployment of 5G networks in many metropolitan areas, new cooperative driving ... [more ▼]

Over the past few years, Connected and Autonomous Vehicles (CAVs) have gained significant research attention. With the recent deployment of 5G networks in many metropolitan areas, new cooperative driving concepts are emerging. One of those is cooperative perception, where vehicles exchange sensory information via a V2X network to maximize their awareness horizon without the need for additional and potentially expensive sensors. The idea is to distribute the processing of the sensor information to find the best trade-off between data transmission and processing time. To experimentally evaluate the performance of cooperative perception schemes is time-consuming and costly due to the expensive hardware it involves. To the best of our knowledge there is to date no open- source simulation framework that allows to transfer realistic sensor data between multiple simulated vehicles via a V2X network. In this work-in-progress paper, we address this issue by proposing an extension of the well-known CARLA open- source simulator for automated driving research. We implement a basic communication channel on top of the existing client-server architecture of CARLA and show how sensor information can be exchanged. Transferring realistic sensor information between multiple vehicles opens up a wide range of experiments to test and evaluate novel approaches to collaborative driving in simulation. [less ▲]

Towards Privacy Preserving Data Centric Super App

in Carvalho Ota, Fernando Kaway; Meira, Jorge Augusto; Frank, Raphaël (Eds.) et al 2020 Mediterranean Communication and Computer Networking Conference, Arona 17-19 June 2020 (2020, September 10)

The number of smartphone users recently surpassed the numbers of desktop users on Internet, and opened up countless development challenges and business opportunities. Not only the fact that the majority ... [more ▼]

The number of smartphone users recently surpassed the numbers of desktop users on Internet, and opened up countless development challenges and business opportunities. Not only the fact that the majority of users are connected using their smartphones, but the number of Internet users in general has popularized the massive use of data-driven applications. In this context, the concept of super apps seems to be the next game-changer for the mobile apps industry, and the challenges related to security and privacy are key aspects for keeping user data safe. Thus, by combining different components for provisioning, authentication, membership and others, we propose a novel framework that enables the creation of a super app using privacy by design principles. [less ▲]

Intrusion detection on robot cameras using spatio-temporal autoencoders: A self-driving car application

in 91st IEEE Vehicular Technology Conference, VTC Spring 2020, Antwerp, Belgium, May 25-28, 2020 (2020)

Robot Operating System (ROS) is becoming more and more important and is used widely by developers and researchers in various domains. One of the most important fields where it is being used is the self ... [more ▼]

Robot Operating System (ROS) is becoming more and more important and is used widely by developers and researchers in various domains. One of the most important fields where it is being used is the self-driving cars industry. However, this framework is far from being totally secure, and the existing security breaches do not have robust solutions. In this paper we focus on the camera vulnerabilities, as it is often the most important source for the environment discovery and the decision-making process. We propose an unsupervised anomaly detection tool for detecting suspicious frames incoming from camera flows. Our solution is based on spatio-temporal autoencoders used to truthfully reconstruct the camera frames and detect abnormal ones by measuring the difference with the input. We test our approach on a real-word dataset, i.e. flows coming from embedded cameras of self-driving cars. Our solution outperforms the existing works on different scenarios. [less ▲]

Evaluation of End-To-End Learning for Autonomous Driving: The Good, the Bad and the Ugly

in 2nd International Conference on Intelligent Autonomous Systems, Singapore, Feb. 28 to Mar. 2, 2019 (2019, March 01)

Visualizing the Learning Progress of Self-Driving Cars

in 21st International Conference on Intelligent Transportation Systems (2018, November 02)

Using Deep Learning to predict lateral and longitudinal vehicle control, i.e. steering, acceleration and braking, is becoming increasingly popular. However, it remains widely unknown why those models ... [more ▼]

Using Deep Learning to predict lateral and longitudinal vehicle control, i.e. steering, acceleration and braking, is becoming increasingly popular. However, it remains widely unknown why those models perform so well. In order for them to become a commercially viable solution, it first needs to be understood why a certain behavior is triggered and how and what those networks learn from human-generated driving data to ensure safety. One research direction is to visualize what the network sees by highlighting regions of an image that influence the outcome of the model. In this vein, we propose a generic visualization method using Attention Heatmaps (AHs) to highlight what a given Convolutional Neural Network (CNN) learns over time. To do so, we rely on a novel occlusion technique to mask different regions of an input image to observe the effect on a predicted steering signal. We then gradually increase the amount of training data and study the effect on the resulting Attention Heatmaps, both in terms of visual focus and temporal behavior. [less ▲]