US 11,797,013 B2
Collision avoidance method and mobile machine using the same
Dejun Guo, Pasadena, CA (US); Kang-Hao Peng, Pasadena, CA (US); Dan Shao, Pasadena, CA (US); Yang Shen, Los Angeles, CA (US); and Huan Tan, Pasadena, CA (US)
Assigned to UBTECH NORTH AMERICA RESEARCH AND DEVELOPMENT CENTER CORP, Pasadena, CA (US); and UBTECH ROBOTICS CORP LTD, Shenzhen (CN)
Filed by UBTECH NORTH AMERICA RESEARCH AND DEVELOPMENT CENTER CORP, Pasadena, CA (US); and UBTECH ROBOTICS CORP LTD, Shenzhen (CN)
Filed on Dec. 25, 2020, as Appl. No. 17/134,219.
Prior Publication US 2022/0206499 A1, Jun. 30, 2022
Int. Cl. G05D 1/02 (2020.01)
CPC G05D 1/0214 (2013.01) [G05D 1/0223 (2013.01); G05D 1/0238 (2013.01); G05D 1/0257 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for a mobile machine to avoid collision with an object, wherein the mobile machine has a plurality of sensors, and the method comprises:
at one or more processors of the mobile machine,
fusing sensor data received from the plurality of sensors to obtain a plurality of data points corresponding to the object;
calculating a closed-form solution of a distance between the mobile machine and each of the plurality of data points;
calculating a maximum allowed velocity of the mobile machine based on the shortest distance between the mobile machine and the plurality of data points and a current velocity of the mobile machine; and
controlling the mobile machine to move according to the maximum allowed velocity;
wherein one of the plurality of sensors is a depth camera; and
the fusing sensor data received from the plurality of sensors to obtain the plurality of data points corresponding to the object comprises:
segmenting the sensor data received from the depth camera to obtain segmented sensor data corresponding to the object;
estimating a velocity of the object based on the segmented sensor data, and predicting a trajectory of the object based on the velocity;
fusing the segmented sensor data and the sensor data received from the other of the plurality of sensors to obtain the plurality of data points corresponding to the object; and
adding a closest point in the predicted trajectory to the plurality of data points.