Methods and apparati for predicting and quantifying threat being experienced by a modeled system

• US 8,543,261 B2
• Filed: 07/15/2010
• Issued: 09/24/2013
• Est. Priority Date: 03/05/2009
• Status: Active Grant

• Current Assignee: Massachusetts Institute of Technology
• Original Assignee: Massachusetts Institute of Technology
• Inventors: Anderson, Sterling J., Peters, Steven C., Iagnemma, Karl D.
• Primary Examiner(s): Tarcza, Thomas
• Assistant Examiner(s): Dunn, Alex C
• Application Number:
  US13/254,746
• Publication Number:
  US 20120109610A1
• Time in Patent Office: 1,167 Days
• Field of Search: 701/3, 701/36, 701/301, 700/44
• US Class Current: 701/3
• CPC Class Codes: B60W 30/09 Taking automatic action to avoid collision, e.g. braking and steering
• CPC Class Codes: B60W 50/0098 Details of control systems ensuring comfort, safety or stability not otherwise provided for
• CPC Class Codes: B60W 50/10 Interpretation of driver requests or demands
• CPC Class Codes: G05D 1/00 Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots drive control systems specially adapted for autonomous road vehicles B60W60/00
• CPC Class Codes: G05D 1/0088 characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours using knowledge based models G06N5/00
• CPC Class Codes: G08G 1/165 for passive traffic, e.g. including static obstacles, trees
• CPC Class Codes: G08G 1/166 for active traffic, e.g. moving vehicles, pedestrians, bikes

Abstract

A method predicts and quantifies the threat posed to a human-operated device based on an optimal device trajectory through a constraint-bounded corridor. A model of the device together with a model of anticipated hazards and the current state of both the device and the hazards are used to iteratively generate an optimal device trajectory through a constraint-bounded corridor or region within state space. Device dynamics are forward-simulated over a time horizon. A method generates a threat assessment metric from the resulting sequence of optimal vehicle states. This threat assessment may be used to devise various types and levels of operator assistance. The human operator can control the device within a safe corridor or region. Threat assessment is based on the nearness of successive optimal trajectory predictions to limits of safe device handling rather than on deviation from a predefined path.

Claims

1. A method for generating a signal that represents an assessment of threat to a mechanical device, operating in an environment, the method comprising the steps of:

a. using a computer control system, using a model of the environment and a model of the device, generating at least one optimal device state trajectory from a current state through a time horizon satisfying constraints imposed by device control inputs, device dynamics, and the position of environmental hazards;

b. using the computer control system, based on at least one generated optimal trajectory prediction, identifying and quantifying threat posed to the device and generating a corresponding threat assessment metric, the at least one generated optimal trajectory prediction being based at least on predicted stability of the device through the time horizon as a function of the at least one optimal device state trajectory; and

c. using the computer control system, based on the threat assessment metric and at least one threat threshold relating to at least one of known, approximated, and desired stability limits and safety margins of the device, generating at least one threat assessment signal.

2. The method of claim 1, the time horizon comprising an infinite horizon.

3. The method of claim 1, the time horizon comprising a finite horizon.

4. The method of claim 1, the model of the environment comprising a changeable model.

5. The method of claim 1, the model of the device comprising an adaptable model.

6. The method of claim 1, the model of the device taking into account non-linear dynamics.

7. The method of claim 1, further comprising the step of, using the computer control system, based on the threat assessment signal, generating a humanly perceptible threat assessment signal.

8. The method of claim 7, the humanly perceptible signal comprising a signal selected from the group consisting of:

an audible signal, a visual signal, a haptic signal, and an olfactory signal.

9. The method of claim 8, the humanly perceptible signal varying in at least one of frequency and amplitude based on the threat assessment metric.

10. The method of claim 1,

a. wherein:

i. the model of the environment comprises a constraint bounded corridor passing through the environment and containing in its interior none of the following;

A. current location of any obstacle; and

B. current location of any environmental hazard; and

ii. the model of the device comprises at least one aspect selected from the group consisting of;

A. a device dynamic constraint;

B. a device input constraint; and

C. a device variable state;

b. the method further comprising the steps of using the computer control system;

i. using the model of the environment and the model of the device, generating constraints on at least one current and predicted device state in an n-dimensional space, where n>

0, to generate an n-dimensional region of operability;

ii. conducting at least one step selected from the group consisting of;

A. generating a sequence of predicted device states that are optimal with respect to an objective function and satisfy input and state constraints; and

B. generating a sequence of device inputs that, when applied to a model of the device, generate a sequence of predicted states that is optimal with respect to an objective function and satisfies input and state constraints;

iii. based on at least one predicted state from the sequence of optimal states, generating the aggregate threat assessment metric; and

iv. comparing the aggregate threat assessment metric to the predetermined threat thresholds and based on the comparison, generating the threat assessment signal.

11. The method of claim 10, further the constraint bounded corridor containing in its interior none of the following:

a. a predicted location of an obstacle; and

b. a predicted location of an environmental hazards.

12. The method of claim 10, the environmental hazard comprising at least one selected from the group consisting of:

terrain features, lane boundaries, road edges, static obstacles, and moving obstacles.

13. The method of claim 10, the device comprising a ground vehicle, the device dynamic and input constraints being selected from the group consisting of:

vehicle position maximum steer angle, minimum steer angle, maximum steer rate, minimum steer rate, maximum acceleration due to braking, minimum acceleration due to braking, maximum acceleration torque, minimum acceleration torque, maximum wheel sideslip angle, maximum tire cornering stiffness and maximum roll angle.

14. The method of claim 10, the step of generating the aggregate threat assessment metric being based on using all of the predicted states of the sequence of optimal states.

15. The method of claim 10, the step of generating the aggregate threat assessment metric being based on at least one of current and past driver performance.

16. The method of claim 10, the step of generating the aggregate threat assessment metric being based on at least one aspect selected from the group consisting of:

situational complexity, situational severity, situational circumstance, risk of collision, risk of instability, and risk of loss of control.

17. The method of claim 10, the device comprising a ground vehicle, the states being selected from the group consisting of:

lateral acceleration, longitudinal wheel slip, wheel sideslip angle, vehicle sideslip angle, vehicle roll angle, lateral load transfer, longitudinal load transfer, and vehicle yaw rate.

18. The method of claim 10, the step of generating the aggregate threat assessment metric comprising basing the generation on a maximum predicted value of at least one state over a prediction horizon.

19. The method of claim 18, the step of basing the generation on a maximum predicted value comprising weighting state predictions that are closer in time to a vehicle state at a chosen time more heavily in the aggregate threat metric than predicted states at different times.

20. The method of claim 10, the step of generating the aggregate threat assessment metric comprising basing the generation on an average predicted value of at least one state over a prediction horizon.

21. The method of claim 20, the step of basing the generation on an average predicted value comprising weighting state predictions that are closer in time to a vehicle state at a chosen time more heavily in the aggregate threat metric than predicted states at different times.

22. The method of claim 10, the step of generating the aggregate threat assessment metric comprising basing the generation on a Root Mean Square (RMS) of predicted value of at least one state over a prediction horizon.

23. The method of claim 22, the step of basing the generation on a Root Mean Square of predicted value comprising weighting state predictions that are closer in time to a vehicle state at a chosen time more heavily in the aggregate threat metric than predicted states at different times.

24. The method of claim 10, the step of generating the aggregate threat assessment metric comprising basing the generation on a maximum predicted value of at least one device control input over a prediction horizon.

25. The method of claim 24, the step of basing the generation on a maximum predicted value comprising weighting input predictions that are closer in time to a vehicle input at a chosen time more heavily in the aggregate threat metric than predicted states at different times.

26. The method of claim 10, the step of generating the aggregate threat assessment metric comprising basing the generation on an average predicted value of at least one device control input over a prediction horizon.

27. The method of claim 26, the step of basing the generation on an average predicted value comprising weighting device control input predictions that are closer in time to a vehicle input at a chosen time more heavily in the aggregate threat metric than predicted inputs at different times.

28. The method of claim 10, the step of generating the aggregate threat assessment metric comprising basing the generation on a Root Mean Square (RMS) of predicted value of at least one device control input over a prediction horizon.

29. The method of claim 28, the step of basing the generation on a Root Mean Square of predicted value comprising weighting device control input predictions that are closer in time to a vehicle input at a chosen time more heavily in the aggregate threat metric than predicted inputs at different times.

30. The method of claim 10, the step of generating the aggregate threat assessment metric comprising determining a threat cost at each prediction step in the optimal sequence using a weighting function of the same form as the objective function used to generate a sequence of at least one optimal device state and optimal device inputs and combining the sequence of determined threat costs to generate the aggregate threat assessment metric.

31. The method of claim 30, the step of generating the aggregate threat assessment metric comprising basing the generation on at least one norm, selected from the group consisting of:

a. a maximum predicted value of at least one state over a prediction horizon;

b. an average predicted value of at least one state over a prediction horizon

c. a Root Mean Square (RMS) of predicted value of at least one state over a prediction horizon;

d. a maximum predicted value of at least one input over a prediction horizon;

e. an average predicted value of at least one input over a prediction horizon

f. a Root Mean Square (RMS) of predicted value of at least one input over a prediction horizon; and

g. a weighted prediction of any of the factors of states and inputs listed in sub-paragraphs a-f of this claim, formed by weighting predictions that are closer in time to a device factor at a chosen time more heavily in the aggregate threat metric than a predicted factor at different times.

32. The method of claim 10, the device comprising a ground vehicle, the predetermined threat thresholds being selected from the group consisting of:

maximum sideslip angle before loss of control, maximum load transfer before wheel liftoff, maximum lateral acceleration before skidding, maximum longitudinal acceleration before skidding, maximum total acceleration before skidding, maximum steer angle before actuator saturation, maximum available acceleration torque, and maximum available braking torque.

33. The method of claim 10, further comprising the step of, using the computer control system, normalizing the threat assessment metric against a predetermined limit.

34. The method of claim 33, the predetermined limit being selected from the group consisting of:

a physical constraint, a dynamic constraint, and a preference.

35. The method of claim 1, further comprising the step of, using the computer control system, based on the threat assessment metric, generating a passive device control assistance.

36. The method of claim 35, the device comprising a ground vehicle, the passive device control assistance comprising at least one selected from the group consisting of:

resistance torques on a steering input member;

resistance torques on a braking input member;

resistance torques on an acceleration input member;

acceleration torques;

braking torques;

priming wheel brakes, preparing safety devices and modifying device properties.

37. The method of claim 1, further comprising the step of, using the computer control system, based on the threat assessment metric, generating an active device control assistance.

38. The method of claim 37, the device comprising a ground vehicle, the active device control assistance comprising at least one selected from the group consisting of:

traction control, adaptive cruise control, yaw stability control, roll stability control, electronic stability control;

lane keeping and obstacle avoidance through active steering control, lane keeping and obstacle avoidance through the imposition of breaking inputs;

lane keeping and obstacle avoidance through the imposition of accelerator inputs.

39. The method of claim 1, further comprising the step of, using the computer control system, based on the threat assessment metric, generating an active device modulation signal.

40. The method of claim 39, the vehicle comprising a ground vehicle, the device modulation signal comprising:

suspension modification.

41. The method of claim 1, the device comprising a ground vehicle.

42. The method of claim 1, the device comprising an aerial vehicle.

43. The method of claim 1, the device comprising a naval vessel.

44. The method of claim 1, the device comprising heavy construction equipment.

45. The method of claim 1, the device comprising a fork lift truck.

46. The method of claim 1, the device comprising a surgical device.

47. The method of claim 1, the device comprising a teleoperated device residing in the environment, which device environment is distant from an environment in which a human device operator resides.

48. A threat assessment apparatus that generates a signal that represents an assessment of threat to a mechanical device, operating in an environment, the apparatus comprising:

a. a processor within which is represented a model of the environment and a model of the device, which processor is configured to generate at least one optimal device state trajectory from a current state through a time horizon satisfying constraints imposed by device controller inputs, device dynamics, and the position of environmental hazards;

b. a processor which is configured to, based on the at least one generated optimal trajectory prediction, identify and quantify threat posed to the device and to generate a corresponding threat assessment metric, the at least one generated optimal trajectory prediction being based at least on predicted stability of the device through the time horizon as a function of the at least one optimal device state trajectory; and

c. a processor which is configured to, based on the threat assessment metric and at least one threat threshold relating to at least one of known, approximated, and desired stability limits and safety margins of the device, generate at least one threat assessment signal.

49. The apparatus of claim 48, further,

a. wherein:

i. the model of the environment comprises a constraint bounded corridor that excludes;

A. current location of any obstacle; and

B. current location of any environmental hazard; and

ii. the model of the device comprises at least one aspect selected from the group consisting of;

A. a device dynamic constraint;

B. a device input constraint; and

C. a device variable state;

b. the apparatus further configured to;

i. using the model of the environment and the model of the device, generate constraints on at least one current and predicted device state in an n-dimensional space, where n>

0, to generate an n-dimensional region of operability;

ii. conduct at least one step selected from the group consisting of;

A. generating a sequence of predicted device states that are optimal with respect to an objective function and satisfy, input and state constraints;

B. generating a sequence of device inputs that, when applied to a model of the device, generate a sequence of predicted states that is optimal with respect to an objective function and satisfies input and state constraints;

iii. based on at least one predicted state of the sequence of states, generate the aggregate threat assessment metric; and

iv. compare the aggregate threat assessment metric to the predetermined threat thresholds and based on the comparison, generate the threat assessment signal.

50. The apparatus of claim 49, further the constraint bounded corridor excluding all locations of at least one, selected from the group consisting of:

a. a predicted location of an obstacle; and

b. a predicted location of an environmental hazards.

51. The apparatus of claim 49, the step of generating the aggregate threat assessment metric being based on using all of the predicted states of the sequence of optimal states.

52. The apparatus of claim 49, the step of generating the aggregate threat assessment metric comprising basing the generation on at least one norm, selected from the group consisting of:

a. a maximum predicted value of at least one state over a prediction horizon;

b. an average predicted value of at least one state over a prediction horizon

c. a Root Mean Square (RMS) of predicted value of at least one state over a prediction horizon;

d. a maximum predicted value of at least one input over a prediction horizon;

e. an average predicted value of at least one input over a prediction horizon

f. a Root Mean Square (RMS) of predicted value of at least one input over a prediction horizon; and

g. a weighted prediction of any of the factors of states and inputs listed in sub-paragraphs a-f of this claim, formed by weighting predictions that are closer in time to a factor at a chosen time more heavily in the aggregate threat metric than predicted factor at different times.

53. The apparatus of claim 49, the step of generating the aggregate threat assessment metric comprising determining a threat cost at each prediction step in the optimal sequence using a weighting function of the same form as the objective function used to generate a sequence of at least one optimal device state and optimal device input and combining the sequence of determined threat costs to generate the aggregate threat assessment metric.

54. The apparatus of claim 52, the step of generating the aggregate threat assessment metric comprising basing the generation on at least one norm, selected from the group consisting of:

a. a maximum predicted value of at least one state over a prediction horizon;

b. an average predicted value of at least one state over a prediction horizon

c. a Root Mean Square (RMS) of predicted value of at least one state over a prediction horizon;

d. a maximum predicted value of at least one input over a prediction horizon;

e. an average predicted value of at least one input over a prediction horizon

f. a Root Mean Square (RMS) of predicted value of at least one input over a prediction horizon; and

g. a weighted prediction of any of the factors of states and inputs listed in sub-paragraphs a-f of this claim, formed by weighting predictions that are closer in time to a factor at a chosen time more heavily in the aggregate threat metric than a predicted factor at different times.