Adaptive and Spatial Aware Embodied Navigation System for Cloud-based Robotics

Project: Research

View graph of relations

Description

As modern urban environments, epitomized by cities like Hong Kong, burgeon in complexity and density, the imperative for advanced navigation systems in service robotics becomes increasingly pronounced. Such navigation systems are crucial for the efficiency and autonomy of robots deployed in service sectors from healthcare to public transport. Current robotic navigation systems predominantly utilize traditional algorithms and sensor inputs. However, these often fall short in densely populated, architecturally intricate environments where dynamic and unpredictable factors become the norm. One of the predominant challenges lies in the robot's ability to adapt in real-time to a myriad of ever-changing scenarios, which current algorithms fail to address efficiently. Innovative integration of cutting-edge navigation approach, physical world rewarding, and cloudbased collective intelligence presents a promising approach, with potential applications far surpassing existing models.In this research project, the PI aims to investigate two novel components: a real-time adaptive learning algorithm and an integrative spatial awareness system. These components are poised to bridge the extant gap in robotic navigation. Our preliminary studies employing these components in campus environments have indicated a substantial increase in navigational accuracy and adaptability in comparison to contemporary systems. By harnessing shared robot-to-robot communication and knowledge, our advanced system facilitates swift environmental adaptation, positioning service robots to not just navigate but also predict and respond preemptively to potential obstacles. Such navigation ability indicates that our system has the potential to be deployed in crowded or high-traffic areas, such as hospitals, warehouses, and public transport hubs in Hong Kong, and has showcased robots navigating with remarkable dexterity, underscoring the system's potential.The outcomes of this research are anticipated to redefine the paradigms of robotic navigation in complex environments, not just in service robots but across various domains requiring intricate navigation. Furthermore, the methodologies and insights garnered from this project could pave the way for broader applications in other AI-driven navigation systems.

Detail(s)

Project number9048304
Grant typeECS
StatusNot started
Effective start/end date1/01/25 → …