Irobot Corporation

Sony Group Corporation

Sony Corporation of America

Vicarious Surgical Inc.

Fabrizio Santini

Joseph M Johnson

Marcus R Williams

Ping-Hong Lu

Peter R Wurman

James MacGlashan

Piyush Khandelwal

Sammy Khalifa

Samuel Barrett

Varun Kompella

Goker Erdogan

A real-time temporal convolution network (RT-TCN) algorithm reuses the output of prior convolution operations in all layers of the network to minimize the computational requirements and memory footprint of a TCN during real-time evaluation. Further, a TCN trained via the fixed-window view, where the TCN is trained using fixed time splices of the input time series, can be executed in real-time continually using RT-TCN.

Computing temporal convolution networks in real time

Systems and methods are used to adapt the coefficients of a proportional-integral-derivative (PID) controller through reinforcement learning. The approach for adapting PID coefficients can include an outer loop of reinforcement learning where the PID coefficients are tuned to changes in the environment and an inner loop of PID control for quickly reacting to changing errors. The outer loop can learn and adapt as the environment changes and be configured to only run at a predetermined frequency, after a given number of steps. The outer loop can use summary statistics about the error terms and any other information sensed about the environment to calculate an observation. This observation can be used to evaluate the next action, for example, by feeding it into a neural network representing the policy. The resulting action is the coefficients of the PID controller and the tunable parameters of things such as the filters.

Methods and systems to adapt PID coefficients through reinforcement learning

A system and method for generating a depth map from image data in a surgical robotic system that employs a robotic subsystem having a camera assembly with first and second cameras for generating image data. The system and method generates based on the image data a plurality of depth maps, and then converts the plurality of depth maps into a single combined depth map having distance data associated therewith. The system and method can then control the camera assembly based on the distance data in the single combined depth map.

System and method for determining depth perception in vivo in a surgical robotic system

A robot includes a body that is movable relative to a surface one or more measurement devices within the body to output information based on an orientation of the body at an initial location on the surface, and a controller within the body to determine an orientation of the body based on the information and to restrict movement of the body to an area by preventing movement of the body beyond a barrier that is based on the orientation of the body and the initial location.

Restricting movement of a mobile robot

Cleaning robots may use floor-type-detection techniques as a trigger for autonomously altering various floor-cleaning characteristics. In some examples, a controller circuit of the robot is configured to determine a flooring type as a function of a signal from a motion sensor indicative of a change in pitch caused by the robot crossing a flooring discontinuity. In some examples, the controller circuit is configured to determine a flooring type based on a power draw signal corresponding to the cleaning head assembly of the robot.

Mobile floor-cleaning robot with floor-type detection

Growing community of inventors

Arlington, MA, United States of America

Fabrizio Santini