Modular autopilot design and development featuring bayesian non-parametric adaptive control

Patent metadata

Field	Value
Publication number	US-10606283-B2
Application number	US-201615751708-A
Country	US
Kind code	B2
Filing date	Aug 15, 2016
Priority date	Aug 13, 2015
Publication date	Mar 31, 2020
Grant date	Mar 31, 2020

Abstract

According to an embodiment, there is provided an onboard integrated computational system for an unmanned aircraft system (“Stabilis” autopilot). This is an integrated suite of hardware, software, and data-to-decisions services that are designed to meet the needs of business and research developers of UAS. Stabilis is designed to accelerate the development of any UAS platform and avionics system; it does so with hardware modularity and software adaptation. The Stabilis offers multiple technological advantages technological advantages including: Plug-and-adapt functionality; Data-to-decisions capability; and, On board parallelization capability.

First claim

What is claimed is:

1. A method of controlling a flight of an aircraft comprising the steps of:
a. selecting at least one control parameter of said aircraft;
b. in a CPU module, performing steps comprising:
(i) determining a current state of said aircraft;
(ii) determining a desired state of said aircraft;
(iii) using said current state and said desired state to estimate a modeling error;
(iv) calculating a tracking error;
(v) using said estimate of modeling error to update a Bayesian Gaussian control process;
(vi) using said updated Bayesian Gaussian control process to calculate an adaptive control factor;
(vii) using at least said adaptive control factor to calculate a pseudocontrol parameter for each of said at least one control parameters;
(viii) calculating a corresponding control surface deflection for each of said at least one pseudocontrol parameters;
(ix) transmitting a signal representative of said corresponding control surface deflection for each of said at least one pseudocontrol parameters to an actuator; and
(x) using said actuator to apply each of said corresponding control surface deflections to the aircraft, thereby adjusting each of said at least one control parameters and controlling the flight of the aircraft.

2. The method of claim 1 , wherein each of said at least one control parameter is selected from the group consisting of position, velocity, and attitude.

3. The method of claim 1 , wherein the step of calculating said adaptive control factor is performed by calculating v ad , where
v ad ( z )= GP ({circumflex over ( m )}( z ), k ( z,z ′))

GP is used to indicate a Bayesian Gaussian process,
({circumflex over (m)}(z) is a mean function of said Gaussian process, and
k(z,z′) is a kernel function of said Gaussian process.

4. The method according to claim 3 , wherein said kernel function k(z,z′) is defined by

k
⁡

(

z
,

z
′

)

=

exp
⁡

(

-



z
,

z
′



2

2
⁢

⁢

μ
2

)

where μ is a mean of said Gaussian process.

5. The method of claim 1 , wherein said aircraft is a fixed wing UAV.

6. The method according to claim 1 , wherein step (b)(viii) comprises the step of:
(1) using at least said adaptive control factor to calculate a pseudocontrol parameter for each of said at least one control parameters using the equation
v=v rm +v pd −v q −v ad ,
where
v is said at least one pseudocontrol parameter,
v rm is the pseudocontrol parameter for the reference model,
v pd is the proportional and derivative control gains,
v q is a robustifying term, and
v ad is said adaptive control factor.

7. The method of claim 1 , wherein said step (b)(x) comprises the steps of:
(i1) calculating an adjustment to each of said control parameters at a point in time via the equation
δ={circumflex over ( f )} −1 ( v,b,t )

where δ is the adjustment to each of said control parameters,
b is a control allocation matrix,
t is said point in time,
v is the pseudocontrol parameter, and
{circumflex over (f)} is an estimate of a system dynamic for the aircraft; and

(i2) adjusting each of said at least one control parameters using δ, thereby controlling the flight of the aircraft.

8. The method of claim 1 wherein said desired state of said aircraft is provided by a pilot.

9. The method according to claim 1 , wherein step (b) is performed a plurality of times.

10. A method of controlling a flight of an aircraft comprising the steps of:
a. selecting at least one control parameter of said aircraft;
b. determining a current state of said aircraft;
c. determining a desired state of said aircraft;
d. in a control CPU module, performing the steps comprising:
(i) using said current state and said desired state to estimate a modeling error;
(ii) calculating a tracking error;
(iii) using at least said estimate of said modeling error to update a Bayesian Gaussian control process;
(iv) using said updated Bayesian Gaussian control process to calculate an adaptive control factor;
(v) using at least said adaptive control factor to calculate a pseudocontrol parameter for each of said at least one control parameters;
(vi) using an estimate of an inverse of a system dynamic of said aircraft to obtain adjustments to each of said at least one control parameters; and
(vii) transmitting a signal representative of said adjustments to each of said at least one control parameters to an actuator; and

(e) using said actuator to apply each of said adjustments to each of said at least one control parameters to produce a control surface deflection of the aircraft, thereby controlling the flight of the aircraft.

11. The method of claim 10 , wherein the step of calculating said adaptive control factor is performed by calculating v ad , where
v ad ( z )= GP ({circumflex over ( m )}( z ), k ( z,z ′)), and

GP is used to indicate a Bayesian Gaussian process,
({circumflex over (m)}(z) is a mean function of said Gaussian process, and
k(z,z′) is a kernel function of said Gaussian process.

12. The method according to claim 11 , wherein said kernel function k(z,z′) is defined by

k
⁡

(

z
,

z
′

)

=

exp
⁡

(

-



z
,

z
′



2

2
⁢

⁢

Assignees

• Univ Oklahoma State

Inventors

• Chowdhary Girish Vinayak
• Stockton Jacob
• Kingravi Hassan

Classifications

• B64C13/16

  actuated automatically, e.g. responsive to gust detectors · CPC title

• G05D1/0825 Primary

  using mathematical models · CPC title

• G06N7/005

  Physics · mapped topic

• B64C39/024

  of the remote controlled vehicle type, i.e. RPV · CPC title

• B64C2201/021

  Operations & Transport · mapped topic