Method and system for controlling a remote vehicle
A system for controlling more than one remote vehicle. The system comprises an operator control unit allowing an operator to receive information from the remote vehicles and send commands to the remote vehicles via a touch-screen interface, the remote vehicles being capable of performing autonomous behaviors using information received from at least one sensor on each remote vehicle. The operator control unit sends commands to the remote vehicles to perform autonomous behaviors in a cooperative effort, such that high-level mission commands entered by the operator cause the remote vehicles to perform more than one autonomous behavior sequentially or concurrently. The system may perform a method for generating obstacle detection information from image data received from one of a time-of-flight sensor and a stereo vision camera sensor.
1. A method of maintaining an occupancy grid map having cells, the method comprising:
receiving image data from a sensor of a remote vehicle, the sensor being one of a time-of-flight sensor and a stereo vision camera sensor, the image data including one or more columns of pixels;
for each pixel of the image data from the sensor:
determining, at a processor operatively connected to the sensor, a direction for the pixel based on a field of view of the sensor; and
combining, at the processor, the direction with depth information of the image data for a potential obstacle in the pixel;
using, at the processor, the combined direction and depth information to plot points, sequentially, for each column of pixels of the image data, each point representing a distance of the potential obstacle from the remote vehicle and a direction of the potential obstacle;
creating, at the processor, one or more best-fit lines from a predetermined number of sequential plotted points;
determining, at the processor, a respective slope of each best-fit line, and determining the existence of the potential obstacle if the slope of the best-fit line is above a threshold slope;
generating, at the processor, a one-dimensional set of values, derived from the best-fit lines, representing the distance to any obstacle detected by the sensor, each value being indexed by its angle with respect to the field of view of the sensor;
updating, at the processor, the occupancy grid map based on the one-dimensional set of values by incrementing a cell of the occupancy grid map, if an obstacle was detected within that cell; and
adding, at the processor, a location of the remote vehicle to the occupancy grid map.
2. The method of claim 1 wherein, a value of the one dimensional set of values is set equal to infinity when an object is not detected in a portion of the image data corresponding to an index of the value.
3. The method of claim 1 , further comprising downsampling the one-dimensional set of values to a predetermined number of values.
4. The method of claim 1 , wherein creating the occupancy grid map from the one-dimensional set of values further comprises decrementing a cell of the occupancy grid map if no obstacle was detected within that cell.
5. The method of claim 1 , wherein more recent depth information is weighted to represent its greater importance in creating the occupancy grid map.
6. The method of claim 5 , wherein the time-of-flight sensor is a SwissRanger sensor or a flash LIDAR sensor.
7. A method for generating obstacle detection information, the method comprising:
receiving image data from a sensor of a remote vehicle, the sensor being one of a time-of-flight sensor and a stereo vision camera sensor, the image data including one or more columns of pixels;
for each pixel of the image data from the sensor:
determining, at a processor operatively connected to the sensor, a direction for the pixel based on a field of view of the sensor; and
combining, at the processor, the direction with a depth information of the image data for a potential obstacle in the pixel;
using, at the processor, the combined direction and depth information to plot points, sequentially, for each column of pixels of the image data, each point representing a distance of the potential obstacle from the remote vehicle and a direction of the potential obstacle;
creating, at the processor, one or more best-fit lines from a predetermined number of sequential plotted points;
determining, at the processor, a respective slope of each best-fit line, and determining the existence of the potential obstacle if the slope of the best-fit line is above a threshold slope; and
generating, at the processor, a one-dimensional set of values, derived from the best-fit lines, representing the distance to any obstacle detected by the sensor, each value being indexed by its angle with respect to the field of view of the sensor.
8. The method of claim 7 , further comprising downsampling the one-dimensional set of values to a predetermined number of values.