Integrating Disciplines for Contextual Robotics

Helpful robots will act based on the physical context, the emotional context and the sociological context that they perceive in real time. Engineers and social scientists from UC San Diego are working together to ensure that the robots of the future will not just do things better, but will be helpful to humanity.

Below are some of the research areas that the Jacobs School of Engineering and the Division of Social Sciences are advancing and integrating in order to develop safe and helpful robotics systems that see, think and do in real time, in the real world, often in close proximity to humans.

CHO Cells
ContextualRobotics.ucsd.edu

SENSING + PERCEPTION

  • Deep learning and statistical analysis of images and video for object detection, scene understanding and context sensing
  • Computational models for rec-ognizing actions and inferring intent and relationships
  • Processing of inputs from real-life applications
  • Sensing, control and optimization algorithms

COGNITION + COORDINATION

  • Distributed decision making and evolution of group behav-ior despite uncertainty and limited communication
  • Embodied Artificial Intelligence
  • Synthetic brain architectures
  • Methods of coupling high-per-formance computing and the Internet of Things with local planning and decision making
  • Conveying ethical and moral imperatives to robot behavior

MOBILITY + MANIPULATION

  • Biologically inspired actuators (limbs) and new materials
  • Robust feedback control mech-anisms for distributed, noisy, unknown environments
  • Models using context to direct safe and appropriate action
  • Coordinated fault-tolerant motion of multiple actuators or vehicles despite limited com-munications and time delays
  • Nano- and micro-robotics

Table of Contents