Autonomous Navigation for Intelligent Vehicles and Robots – NAVIR

The NAVIR project is being led by HawAI.tech in collaboration with the Chroma team at Inria, with the aim of developing, demonstrating, and promoting the use of probabilistic approaches in mobile robotics. To do this, the research team will combine its expertise in perception and embedded navigation with the industrial company's hardware acceleration capabilities.

HawAI.tech is a deeptech startup specializing in accelerating hardware performance to facilitate the advancement of a new form of AI, namely probabilistic AI. By combining data-driven learning systems with expert systems, this AI intends to address a variety of socio-economic concerns such as efficiency, transparency, and responsibility. The company's technology has already produced convincing results in the fields of recognition and recommendation. In the face of unpredictable and imperfect real-world settings, the company provides innovative ways that produce cutting-edge results. The MOBILEX project objectives are precisely aligned with the application criteria that the company's technology addresses, such as the consideration of an unpredictable environment, decision support, and dynamic data analysis.

The Chroma team develops generic probabilistic models and software that generates, fuse, filter, analyse, and exploit probabilistic occupation grids. The system has been meticulously fine-tuned to leverage the power of graphics cards and is currently embedded in numerous demonstrators using GPU cards like Nvidia Jetson TegraX2 and AGX. It excels in handling intricate, ever-changing, unstructured, and uncertain environments, thanks to its utilization of comprehensive and dense representations. In addition, its capacity to account for uncertainties during processing and data analysis, coupled with its seamless integration of information from diverse sensors, further enhances its effectiveness.

The project consists of three phases aligned with the challenge events. Our goal is to deploy and enhance the existing tested methods from Inria to meet the challenge's specific requirements. We will also leverage the HawAI.tech platform to accelerate computations and improve performance. This progressive approach allows us to adapt and optimize our solutions throughout the project. This project aligns with HawAI.tech's strategy to enter the mobility market and transfer technology from Inria.

Creating and operating an experimentation and demonstration platform is well-suited to support the technological advancements of the Chroma team. It will strengthen their scientific output, foster collaborations with industry partners, and increase their visibility and appeal.

Furthermore, the platform aims to demonstrate the broader scientific and industrial significance of AI. It showcases the transformative power of AI, which goes beyond solely relying on data and instead incorporates human models and expertise to drive innovation.