Non linear predictive model for haptic waveform generation
A system includes a sequencer that divides a reference waveform into reference sequences, a sequence adjuster, and a model that models non-linearities of a haptic rendering signal chain that includes a haptic transducer load and a driver to the load. For each reference sequence: the sequence adjuster transforms the reference sequence into a test sequence using one or more parameters (e.g., changes reference sequence amplitude and/or period), the model generates an output in response to the test sequence, an error signal is generated that measures a difference between the output and the reference sequence, and if the error signal is above a threshold the parameters are adjusted based on the error signal. These operations are repeated until the error signal is below the threshold in which case the test sequence becomes a selected sequence, which is then sent to the haptic rendering signal chain.
1. A system, comprising:
a sequencer that divides a reference waveform into reference sequences;
a sequence adjuster;
a model of a haptic rendering signal chain, wherein the haptic rendering signal chain includes a haptic transducer load and a driver to the load;
wherein for each reference sequence of the reference sequences:
(a) the sequence adjuster transforms the reference sequence into a test sequence using one or more parameters;
(b) the model generates an output in response to the test sequence;
(c) an error signal is generated that measures a difference between the output and the reference sequence;
(d) if the error signal is above a threshold:
the one or more parameters are adjusted based on the error signal;
operations (a) through (d) are repeated until the error signal is below the threshold in which case the test sequence becomes a selected sequence; and
the selected sequence is sent to the haptic rendering signal chain;
wherein the haptic rendering signal chain further includes a sensor that senses signals comprising a load current and/or a load voltage of the load;
wherein the model of the haptic rendering signal chain includes a model of the sensor; and
an estimator that transforms the outputs of the model of the sensor into a signal that estimates a modeled haptic effect rendered by the model of the haptic rendering signal chain in response to each of the test sequences.
2. The system of claim 1 ,
wherein the model models non-linearities of the haptic rendering signal chain that include one or more of the following non-linearities:
driver clipping;
load current distortion;
load voltage distortion;
load excursion limiting;
force factor distortion; and
non-linear spring compression.
3. The system of claim 1 ,
wherein the reference waveform is one of:
an acceleration profile;
a velocity profile;
a position profile;
a driver voltage profile; and
a driver current profile.
4. The system of claim 1 , further comprising:
wherein the load renders a haptic effect in response to the selected sequences;
wherein the haptic effect is measurable as an acceleration, velocity or position of a mass of the load; and
wherein the signal that estimates the modeled haptic effect is a representation of one of:
a back electromotive force (emf) of the load;
a position of a mass of the load;
a velocity of a mass of the load; and
an acceleration of a mass of the load.
5. The system of claim 1 ,
wherein the haptic rendering signal chain further includes at least one of the following:
an equalizer or other linear filter coupled to an input of the driver that compensates for dynamics of the haptic transducer load and driver; and
a controller coupled to an input of the driver that includes a feedback input based on the one or more sense signals of a sensor of the haptic rendering signal chain that senses signals comprising a load current and/or a load voltage of the load.
6. The system of claim 1 , further comprising:
a waveform authoring system that compresses the reference waveform to utilize a full range of the driver.
7. The system of claim 1 ,
wherein the driver is a class-D amplifier having a class-H control waveform;
wherein the class-H control waveform:
is a parameter of the sequence adjuster; or
is computed a posteriori from a voltage requested by the model of the driver.
8. The system of claim 1 ,
wherein one or more parameters of the model of the haptic rendering signal chain are updated from an on-line or off-line load model estimator that estimates an impedance of the load;
wherein the load is an electromechanical load; and
wherein the load impedance parameters comprise one or more of:
a resonant frequency of the electromechanical load;
a bandwidth of the electromechanical load;
a resistance and inductance of a coil of the electromechanical load; and
a resonant amplitude of a capacitive-inductive tank of the electromechanical load;
wherein the haptic rendering signal chain further includes an equalizer or a closed-loop controller; and
wherein models of the load and/or the equalizer or the closed-loop controller are updated from the load model estimator.
9. The system of claim 1 , wherein the reference waveform is generated by a real-time process and streamed to the system.
10. The system of claim 1 ,
wherein the system comprises an integrated circuit that:
comprises the haptic rendering signal chain;
simulates the model in real-time to generate the test sequences; and
sends the selected sequences to the haptic rendering signal chain; and
wherein simulation of the model is performed by a microcontroller unit (MCU), a digital signal processor (DSP), and/or a dedicated signal processing hardware block.
11. The system of claim 1 ,
wherein the load is an electromechanical load.
12. A method for rendering a haptic effect by a haptic rendering signal chain that includes a haptic transducer load and a driver to the load, comprising:
dividing a reference waveform into reference sequences;
for each reference sequence of the reference sequences:
(a) transforming the reference sequence into a test sequence using one or more parameters;
(b) generating, by the model, an output in response to the test sequence;
(c) generating an error signal that measures a difference between the output and the reference sequence;
(d) if the error signal is above a threshold:
adjusting the one or more parameters based on the error signal;
repeating operations (a) through (d) until the error signal is below the threshold in which case the test sequence becomes a selected sequence; and
sending the selected sequence to the haptic rendering signal chain;
wherein the haptic rendering signal chain further includes a sensor that senses signals comprising a load current and/or a load voltage of the load;
wherein the model of the haptic rendering signal chain includes a model of the sensor; and
transforming outputs of the model of the sensor into a signal that estimates a modeled haptic effect rendered by the model of the haptic rendering signal chain in response to each of the test sequences.
13. The method of claim 12 ,
wherein the model models non-linearities of the haptic rendering signal chain that include one or more of the following non-linearities:
driver clipping;
load current distortion;
load voltage distortion;
load excursion limiting;
force factor distortion; and
non-linear spring compression.
14. The method of claim 12 ,
wherein the reference waveform is one of:
an acceleration profile;
a velocity profile;
a position profile;
a driver voltage profile; and
a driver current profile.
15. The method of claim 12 , further comprising:
wherein the load renders a haptic effect in response to the selected sequences;
wherein the haptic effect is measurable as an acceleration, velocity or position of a mass of the load; and
wherein the signal that estimates the modeled haptic effect is a representation of one of:
a back electromotive force (emf) of the load;
a position of a mass of the load;
a velocity of a mass of the load; and
an acceleration of a mass of the load.
16. The method of claim 12 ,
wherein the haptic rendering signal chain further includes at least one of the following:
an equalizer or other linear filter coupled to an input of the driver that compensates for dynamics of the haptic transducer load and driver; and
a controller coupled to an input of the driver that includes a feedback input based on the one or more sense signals of a sensor of the haptic rendering signal chain that senses signals comprising a load current and/or a load voltage of the load.
17. The method of claim 12 , further comprising:
compressing the reference waveform to utilize a full range of the driver.
18. The method of claim 12 ,
wherein the driver is a class-D amplifier having a class-H control waveform;
wherein the class-H control waveform:
is a parameter of the sequence adjuster; or
is computed a posteriori from a voltage requested by the model of the driver.
19. The method of claim 12 ,
wherein one or more parameters of the model of the haptic rendering signal chain are updated from an on-line or off-line load model estimator that estimates an impedance of the load;
wherein the load is an electromechanical load; and
wherein the load impedance parameters comprise one or more of:
a resonant frequency of the electromechanical load;
a bandwidth of the electromechanical load;
a resistance and inductance of a coil of the electromechanical load; and
a resonant amplitude of a capacitive-inductive tank of the electromechanical load;
wherein the haptic rendering signal chain further includes an equalizer or a closed-loop controller; and
updating, from the load model estimator, models of the load and/or the equalizer or the closed-loop controller.
20. The method of claim 12 ,
wherein the reference waveform is generated by a real-time process and streamed to the system.
21. The method of claim 12 ,
wherein the load is an electromechanical load.