Unkown

• US 20150149000A1
• Filed: 07/03/2014
• Published: 05/28/2015
• Est. Priority Date: 01/04/2013
• Status: Active Grant

• Current Assignee: Parrot Drones
• Original Assignee: Parrot SA
• Inventors: RISCHMULLER, MICHAEL, CHEVALLEY, LAURE, CALLOU, FRANCOIS, CALDICHOURY, ETIENNE
• Granted Patent: US 9,387,927 B2
• Time in Patent Office: Days
• US Class Current: 701/7
• CPC Class Codes: B64C 39/024 of the remote controlled vehicle type, i.e. RPV
• CPC Class Codes: B64U 10/10 Rotorcrafts
• CPC Class Codes: B64U 10/14 with four distinct rotor axes, e.g. quadcopters
• CPC Class Codes: B64U 2201/10 autonomous, i.e. by navigating independently from ground or air stations, e.g. by using inertial navigation systems [INS]
• CPC Class Codes: G05D 1/102 specially adapted for vertical take-off of aircraft

Abstract

The drone comprises: a vertical-view camera (132) pointing downward to pick up images of a scene of the ground overflown by the drone; gyrometer, magnetometer and accelerometer sensors (176); and an altimeter (174). Navigation means determine position coordinates (X, Y, Z) of the drone in an absolute coordinate system linked to the ground. These means are autonomous, operating without reception of external signals. They include image analysis means, adapted to derive a position signal from an analysis of known predetermined patterns (210), present in the scene picked up by the camera, and they implement a predictive-filter estimator (172) incorporating a representation of a dynamic model of the drone, with as an input the position signal, a horizontal speed signal, linear and rotational acceleration signals, and an altitude signal.

Claims

1. A rotary-wing drone (10) with multiple rotors (12) driven by respective motors selectively controlled by application of differentiated motor commands to pilot the drone in altitude and speed, comprising:

a plurality of sensors adapted to deliver respective drone-state signals, such sensors comprising;

a vertical-view camera (16, 132) pointing downward, adapted to pick up images of a scene of the ground (20) overflown by the drone and to produce a horizontal speed signal, derived from an analysis of the displacement of the picked-up scene from one image to the following;

gyrometer (102), magnetometer (116) and accelerometer (114) sensors, adapted to deliver linear and rotational acceleration signals; and

an altimeter (22, 152-158), adapted to deliver a signal of drone altitude relative to the overflown ground; and

a predictive-filter estimator incorporating a representation of a dynamic model of the drone, this filter being adapted to perform a prediction of at least one state value of the drone based on the motor commands and to periodically readjust this prediction as a function of the signals delivered by the sensors,

wherein the drone further comprises navigation means adapted to determine position coordinates of the drone in an absolute coordinate system (NED) linked to the ground, and in that these navigation means;

are autonomous means adapted to operate without reception of signals external to the drone;

comprise image analysis means (160), adapted to derive a position signal from an analysis of known predetermined patterns (210), present in the scene picked up by the camera;

implement the predictive-filter estimator (172) at the input of which are applied;

the position signal;

the horizontal speed signal;

the linear and rotational acceleration signals; and

the altitude signal.

2. The drone of claim 1, wherein the predictive filter (172) is an eight-state filter, these states comprising:

two horizontal components (V_D/SX, V_D/SY) of the speed of displacement of the drone relative to the ground, expressed in a coordinate system (u,v,w) linked to the drone;

two horizontal components (V_A/SX, V_A/SY) of the speed of air relative to the ground, expressed in the absolute coordinate system (NED) linked to the ground;

two horizontal components (B_X, B_Y) of the drone accelerometer bias; and

two horizontal coordinates (X_NED, Y_NED) of the drone position in the absolute coordinate system (NED) linked to the drone.

3. The drone of claim 2, wherein the predictive filter (172) is a Kalman predictive filter.

4. The drone of claim 1, wherein the altimeter also comprises a predictive filter, adapted to operate a prediction of altitude of the drone and to periodically readjust this prediction as a function of:

the signals delivered by at least one between a telemeter sensor (154) and a barometer sensor (156); and

a vertical coordinate signal (Z_NED) of the drone position, derived from said analysis of the known predetermined patterns (210) present in the scene picked up by the camera.

5. The drone of claim 1, wherein the means for analyzing the known predetermined patterns, present in the scene picked up by the camera comprise:

prefiltering means, adapted to extract from the image delivered by the camera the pixels corresponding to the edges of the patterns (210) and of sub-patterns (212a-212d), and to deliver the position thereof and the orientation of the contour (214, 216) of each pattern and sub-pattern;

colorimetric means, adapted to determine the proper color of each sub-pattern;

recognition means, adapted to identify each pattern based on the proper colors of each of the its sub-patterns; and

position determination means, adapted to determine the position coordinates (X_NED, Y_NED, Z_NED) of the drone in the absolute coordinate system (NED) linked to the ground as a function of the patterns identified by the recognition means.

6. The drone of claim 5, wherein the position determination means are means adapted to evaluate a distance between the drone and a pattern as a function i) of the size of the pattern in the image picked up by the camera, ii) of the known real size of the pattern and iii) of the geometric characteristics of the optical system of the camera.

7. The drone of claim 5, wherein the position determination means are means adapted to evaluate a distance between the drone and a plurality of patterns (210a-210d) by i) determining for each pattern a straight line (242a-242d) linking each pattern to the drone and ii) determining the position by applying to these straight lines a least-square minimization algorithm.

8. A unit comprising:

a drone (10) according to any one of claims 1 to 7, and

carpet (200) carrying an array of said predetermined patterns (210), each of these patterns being uniquely identified by proper visual characteristics.

9. The unit of claim 8, wherein said proper visual characteristics comprise for each of the patterns a plurality of colored sub-patterns (212a-212d), the color of each sub-pattern being chosen in a predetermined series of colors (R, G, B, Y, M), and the colors of the sub-patterns of a same pattern being uniquely chosen, and invariant by rotation, for all the patterns of the carpet.