Intention prediction (IP) is a challenging task for intelligent vehicle’s perception systems. IP provides the likelihood, or probability, of a target vehicle to perform a maneuver subjected to a finite set of possibi...
Intention prediction (IP) is a challenging task for intelligent vehicle’s perception systems. IP provides the likelihood, or probability, of a target vehicle to perform a maneuver subjected to a finite set of possibilities. There are many factors that influence the decision-making process of a driver, which should be considered in a prediction framework. In addition, the lack of labeled large-scale dataset with maneuver annotation imposes another challenge to the task. This paper proposes an Interaction-aware Maneuver Prediction framework, called IAMP, using interaction graphs to extract complex interaction features from traffic scenes. In addition, we explored a semi-supervised approach called Noisy Student to take advantage of unlabeled data in the training step. Experimental results show relevant improvement when using unlabeled data, increasing the average performance of a classifier by 7.17% of accuracy. Moreover, this approach also made it possible to obtain an intention predictor with similar results to a classifier., even when using a shorter observation horizon.
This paper addresses the problem of detecting humans in RGB and Thermal (long-wave IR) images taken by cameras mounted onboard a mobile robot. Human/Pedestrian detection is currently one of the most pertinent object d...
This paper addresses the problem of detecting humans in RGB and Thermal (long-wave IR) images taken by cameras mounted onboard a mobile robot. Human/Pedestrian detection is currently one of the most pertinent object detection problems, mainly due to safety concerns in autonomous vehicles. The majority of approaches apply deep-learning techniques based solely on RGB images. However, they have a few shortcomings, namely that during foggy weather, nighttime, and low-light scenarios, these images may not contain sufficient information. To address these issues, this work studies the use of thermal cameras as a complementary source of information for human detection in indoor and outdoor environments. The proposed approach uses YOLOv5 to detect pedestrians in both thermal and RGB images. Moreover, the different modalities are combined using early and late fusion techniques. Evaluation of the proposed approach is carried out in the FLIR Aligned dataset and in a new in-house dataset. Results indicate that the use of fusion techniques highlights a promising way to improve the overall performance in this application domain.
With the rapid development of edge computing, industrial automation has evolved a two-layer distributed computing architecture, including local controllers communicating through wired networks and edge controllers com...
详细信息
Safety critical systems are typically subjected to hazard analysis before commissioning to identify and analyse potentially hazardous system states that may arise during operation. Currently, hazard analysis is mainly...
Safety critical systems are typically subjected to hazard analysis before commissioning to identify and analyse potentially hazardous system states that may arise during operation. Currently, hazard analysis is mainly based on human reasoning, past experiences, and simple tools such as checklists and spreadsheets. Increasing system complexity makes such approaches decreasingly suitable. Furthermore, testing-based hazard analysis is often not suitable due to high costs or dangers of physical faults. A remedy for this are model-based hazard analysis methods, which either rely on formal models or on simulation models, each with their own benefits and drawbacks. This paper proposes a two-layer approach that combines the benefits of exhaustive analysis using formal methods with detailed analysis using simulation. Unsafe behaviours that lead to unsafe states are first synthesised from a formal model of the system using Supervisory Control Theory. The result is then input to the simulation where detailed analyses using domain-specific risk metrics are performed. Though the presented approach is generally applicable, this paper demonstrates the benefits of the approach on an industrial human-robot collaboration system.
This paper addresses the task of predicting the behavior of traffic participants, which involves complexities such as road geometry and agent interactions. To overcome these challenges, this paper presents a novel fra...
This paper addresses the task of predicting the behavior of traffic participants, which involves complexities such as road geometry and agent interactions. To overcome these challenges, this paper presents a novel framework called AIMP (Attention-based Interaction-aware Maneuver Prediction). AIMP utilizes interaction graphs to extract intricate interaction features from traffic scenes. The framework incorporates a Gated Mixture-of-Experts Attention Mechanism, which combines information from road geometry, interaction patterns, and motion dynamics. This fusion process also considers prior maneuver intention estimations, enhancing both explainability and informativeness. Experimental results highlight a performance enhancement (approximately 2% ∼ 9% of accuracy) of the proposed AIMP framework compared to alternative fusion methods.
The prioritized control problem is a process to find a control strategy for a dynamical system with prioritized multiple outputs, so that it can operate outside its nonsingular domain. Singularity typically leads to i...
The prioritized control problem is a process to find a control strategy for a dynamical system with prioritized multiple outputs, so that it can operate outside its nonsingular domain. Singularity typically leads to imperfect inversion in the prioritized control problem, which in turn results in imperfect input-output feedback linearization. In this paper, we propose a method based on the Kalman-Yakubovich-Popov lemma that compensates nonlinear feedback terms caused by the imperfect inversion of the prioritized control problem. In order to realize this idea, we prove existence of a feedback gain matrix that gives a strictly positive real transfer function whose output matrix is identical to the feedback gain matrix. Our proof is constructive so that a set of such matrices can be found. Also, we provide a numerical approach that gives a larger set of feedback gain matrices and validate the result with numerical examples.
Under certain conditions, the benefit of three-dimensional virtual environments during worker training is unquestionable. When demonstrating the components of special mechanical equipment and training firefighters for...
详细信息
The technological development of the 21st century entails the methodological renewal of public education and higher education, and the application of educational innovations. As a result of this modernisation, educati...
详细信息
For automated, driverless rail transportation applications in open environments, Artificial Intelligence (AI)-based methods are gaining importance, especially in computer vision and perception tasks. The safe operatio...
详细信息
ISBN:
(数字)9798331505929
ISBN:
(纸本)9798331505936
For automated, driverless rail transportation applications in open environments, Artificial Intelligence (AI)-based methods are gaining importance, especially in computer vision and perception tasks. The safe operation of complex automated systems requires validation processes. For this purpose, the concept of Operational Design Domains (ODDs), driven by recent developments in the automotive industry, is gaining momentum, allowing to describe different aspects of operating conditions as scenes and scenarios. With regard to safety and authorization using AI-based vision systems, data coverage is needed, which can be enhanced by employing virtual reality in different forms. The creation of virtual scenes and sensor models allows the generation of synthetic sensor data and metadata that can be used as a database for the training of the vision system.
Robotic Hand glove is one of the most commonly used technique in the rehabilitation systems. In this paper, we developed a robotic hand system with a proposed sensing mechanism-based AI algorithm, which can acquire gr...
详细信息
暂无评论