AI_RIDE – Artificial Intelligence – driven RIding Distributed Eye
Idea and relation to VEDLIoT
The AI-RIDE Project targets the development of an interactive and accelerated AI-driven framework for Practical Driving Courses and Driving Licence Exams. The framework will make use of the VEDLIoT hardware platforms and the software toolchain to enable specific augmented reality available at the driver’s helmet interactive screen, along with an innovative multiparameter AI-assisted telemetry system able to compute test scores and outcome, useful for human-neutral auditability of Driving Licence Exams. The distributed AI system available at the Track Testbed will be able to perform driving behavior classifications and will suggest specific improvements, targeting body posture, vehicle position, and trajectory. Besides different innovations, actual eye pupil tracking and its optimal placement will be evaluated as a significant key performance parameter. Finally, the project will target the creation of a large dataset for driving tests classification of key performance parameters. The system is envisioned to have a relevant impact for all the certification, driving licence operators and regulator entities. An interactive demonstration will be organized at the Track Testbed Premises located in Pontedera (Pisa, Italy) to show the app proof-of-concept of the envisioned use cases during real riding test sessions.
AI-RIDE will contribute to the VEDLIoT ecosystem extension with mainly 3 aspects:
- The VEDLIoT hardware platform will be explored for the system and testbed development.
- Distributed AI will be elaborated by the VEDLIoT toolchain.
- A specific WP (WP2) will be dedicated to the identification of the most suitable architecture compliant with the VEDLIoT platform. Specific efforts in Task 2.1 (requirements), Task 2.2 (architectural strategies) and Task 3.3 (modules and selected techniques) will identify and implement the most suitable set of techniques for each AI-RIDE component with proactive collaboration with the VEDLIoT team.
- Defining the AI-RIDE requirements of the application, the system, the network, the distributed AI, and the sensors required to run AR/VR feedback.
- Defining the AI-RIDE application architecture, the functional relationship with the different sub-system (sensors, network terminals).
- Cover the Track Testbed location setup of the VEDLIoT hardware and software.
- Studying and performing UC1, by planning, recording and classifying the reference AI-RIDE dataset.
- Studying, evaluating, selecting and deploying the hardware and software components required for the AI-RIDE use cases.
- Designing and implementing the AI models for UC2 and UC3 and the different processing modules for the distributed VEDLioT platform that will be installed in the Track Testbed.
- Deploying the AI-RIDE application, including a Graphic User Interface for both instructors and practitioners.
- Integrating, deploying and validating the AI-RIDE software and applications using continuous integration methodology and create automated tests to validate the proper integration of the software developed in WP3.
- Performing the final full demonstration of the AI-RIDE solution in the Track Testbed. The demonstration will include a co-located workshop open to VEDLIoT partners, research associations, automotive consortium and industries, driver license operators.
The AI-RIDE project proposes the adoption of an accelerated, online and embedded Artificial Intelligence framework in the context of motorcycle rider training, particularly targeting the Practical Driving Courses (PDC) and Driving Licence Exam (DLE) sessions verification tools. The project will target a disruptive innovation step in the context of driving learning techniques, significantly going beyond the state of the art of the current instruments used in the PDC and DLE ecosystem.
The VEDLIoT hardware platform and middleware will be the key driver to enable the AI-RIDE platform. Thanks to the disaggregated IoT and sensors-based network architecture, the distributed algorithm approach, high speed networking and embedded online AI at the edge, the AI-RIDE application will be capable of performing a number of innovative functions supporting the whole driving learning ecosystem (i.e., practitioners, licence candidates, instructors, commissioners). The VEDLIoT Far Edge Computing and/or Near Edge Computing platforms, deployed at the motorcycle track facilities, are foreseen as the best candidate platforms guaranteeing adequate online computing capabilities, along with the low latency requirements needed for online assistance. For the scope of AI-RIDE, video and image processing will be required. Thus, the required VEDLIoT hardware is expected to include CPU and GPU platforms to perform data fusion from different sources (cameras and wearables) and implement augmented reality under the form of optimal vision targets during a drive test session. For the specific use cases and the requirements of the application, the Project will rely on VEDLIoT hardware through the loan service support of the open Call. The hardware platforms will be installed and tested locally to achieve low latency performance during driving test sessions.
The short term goal of the AI-RIDE project is to design and validate automated and reliable solutions for AI-accelerated training and testing of motorcycle practical driving courses and driving license exam sessions.
The final goal of the AI-RIDE Project is to increase road traffic safety, improve the driving experience and training process, and introduce ecological driving behaviors in all traffic conditions and scenarios. Such an ambitious goal can be achieved only through a gradual approach, where specific concrete objectives and targeted scenarios are addressed in a prioritized way, guaranteeing a return of investment in the short/medium term. Having a strong short term business plan is indeed crucial in the process of defining a new product and service. Another goal is to disseminate results to the Licence Exams regulatory bodies and institutions (EU, Italian Ministry) to suggest innovative and human-neutral methods for Driving Licence certifications.
The medium term goal of the project is the deployment of an advanced IoT/Edge Pilot testbed for automotive-oriented and smart city apps/vertical industries, open for future research projects.