lecture 3: kinesthetic haptic devices: sensors and actuators · lecture 3: kinesthetic haptic ......

ME 327: Design and Control of Haptic Systems Autumn 2018

Lecture 3:Kinesthetic haptic devices:

Sensors and Actuators

Allison M. OkamuraStanford University

Stanford University ME 327: Design and Control of Haptic Systems © Allison M. Okamura, 2018

pay lab materials fee $50 check made out to Stanford University, by class time Thursday 9/18

do Assignment #1 http://web.stanford.edu/class/me327/#c3do your 3D printing as early as possible!

two optional seminars this week Wednesday @ 1:30 pm MER 203, Friday @ 12:30 pm @ Gates B1

see details on piazza

bring your laptop w/USB on Thursday so you can test communication with the Hapkit Board

http://web.stanford.edu/class/me327/#c3

Office Hours for this week• Allison: Tuesday (today) 4:30-6 pm in 520-145

• Nathan: Wednesday (tomorrow) 3-5 pm in 520-145

• Jake: Wednesday (tomorrow) 5-6 pm in 520-145

• 520-145 (D'Arbeloff Teaching Lab) schedule is online

• You can also post questions to piazza

• You can also email us for an appointment (please email all three of us and be sure to state your available times for the next couple of days)

sensors


sensor types

• magnetic

• optical

• acoustic

• inertial

• mechanical(our focus, since these are the sensors typically integrated with the actuator in kinesthetic haptic devices)


mechanical: Faro arm

optical: Polaris, NDI

magnetic: TrakStar, Ascension

optical: Microsoft Kinect acoustic: ultrasonic proximity sensor, BiF

inertial: wearable IMU, MotionNode

mechanical trackers

• ground-based linkages most commonly used

• joint position sensors

– digital: optical encoders are most common

– analog: magnetic sensors and potentiometers are most common


optical encoders

• how do they work?

–A focused beam of light aimed at a matched photodetector is interrupted periodically by a coded pattern on a disk

–Produces a number of pulses per revolution (Lots of pulses = high cost)

• quantization problems at low speeds

• absolute vs. referential


optical encoders• phase-quadrature encoder

• 2 channels, 90° out of phase

–allows sensing of direction of rotation

–4-fold increase in resolution


magnetoresistive angle sensors

• magnetoresistive materials change their electrical resistance when an external magnetic field is applied

• the resistance depends on the angle between the magnetization vector of the ferromagnetic material and the direction of current flow (resistance is largest if they are parallel)

• often 4 sensors are connected in a Wheatstone bridge configuration (similar to strain gages)


MR sensor

integratedcircuit

KMA199E, NXP

KMA199E, NXP

Hall-Effect SensorsHow do they work?

a small transverse voltage is generated across a current-carrying conductor in the presence of a magnetic field

(Discovery made in 1879, but not useful until the advent of

semiconductor technology.)


Hall-Effect Sensors

• amount of voltage output related to the strength of magnetic field passing through.

• linear over small range of motion (need to be calibrated)

• affected by temperature, other magnetic objects in the environments


potentiometers


measuring velocity

• differentiate position

– advantage: use same sensor as position sensor

– disadvantage: get noisy signal

• alternative

– for encoders, measure time between ticks


discrete differentiation• many different methods

• simple example:

– average 20 readings = P1

– average next 20 readings = P2

– where t is the the period of the servo loop

• differentiation increases noise

• usually need to filter

0 2 4 6 8 10 12012345678910

time

position

0 2 4 6 8 10 120

0.2

0.4

0.6

0.8

1

1.2

1.4

timevelocity


position/velocity filtering• one example is the simple infinite impulse

response filter

• pseudocode for real-time filtering: new_value = read_from_sensor()filtered_value = a*new_value + (1-a)*old_valueold_value = filtered_value


// Return RC low-pass filter output samples, given input samples, // time interval dt, and time constant RC function lowpass(real[0..n] x, real dt, real RC) var real[0..n] y var real α := dt / (RC + dt) y[0] := x[0] for i from 1 to n y[i] := α * x[i] + (1-α) * y[i-1] return y

time-between-ticks

• encoders fare poorly at slow velocities

– there may be very few ticks during a single servo loop

• instead, some specialized data acquisition boards use a special chip that measures time between ticks

– fares poorly at high velocities


discussion

what are the advantages/disadvantages for haptics of the three sensor types discussed

(optical encoders, magnetoresistive angle sensors, and potentiometers)?

are there any common types of sensors you find missing from this discussion?


actuators


actuator types

Burdea


PM DC brushed motors

• rotating armature with coil windingsis caused to rotaterelative to apermanent magnet

• current is transmitted through brushes to armature, and is constantly switched so that the armature magnetic field remains fixed.


DC motor components


DC motor terms• cogging/torque ripple

– tendency for torque output to ripple as the brushes transfer power

• friction/damping

–caused by bearings, brushes, and eddy currents

• stall torque

–max torque delivered by motor when operated continuously without cooling


torque ripple


http://www.maxonmotorusa.com/

motor equations• torque constant

• speed constant

• dynamic equations


⌧ = kT ikT

v = Ldi

dt+Ri+ vemf

vemf = kv q̇vemf

mq̈ + bq̇ = ⌧

motor amplifier types


voltage amplifier(voltage controlled voltage source VCVS)

indirectly controls currentcurrent depends on ???

less expensive (???)

ardumotor shieldhttps://www.sparkfun.com/products/9815

based on L298 H-bridge

current amplifier(voltage controlled current source VCCS)

directly controls currentcurrent = torque (good!)

expensive

OPA544+

-

2 Ohm

Vin

DCmotor

V+

V-

Circuit board design byLouis Whitcomb, JHU

pulse width modulation


useful if you do not have a D/A converterto send analog signals to the motor circuit

switching frequency must be much faster than the mechanical dynamics of the system

assumes that the average signal is a constant signal

duty cycle is the proportion of on time to the period

http://www.barrgroup.com/

transmission• Transfers/amplifies force/torque from motor

• You don’t want to feel or see the effects of the transmission!

• Types:

– gears

– belts/pulleys

– capstan drive

– none (direct drive)Stanford University ME 327: Design and Control of Haptic Systems © Allison M. Okamura, 2018

capstan drive

Katherine Kuchenbecker


high transmitted force, low transmitted friction

Springer Handbook of Robotics

Siciliano, Khatib (Eds.) · ©Springer 20081

726 Part D Manipulation and Interfaces

Many haptic applications, such as the rendering ofvirtual environments with damping ( in which force isproportional to velocity), require velocity measurement.Velocity is typically obtained by numerical differentia-tion of the position signal obtained by an encoder. Analgorithm for velocity estimation must be selected whichis free of noise but minimizes phase lag at the frequen-cies of interest [30.36]. Thus, an alternative method is touse specialized hardware that measures the time betweenencoder ticks in order to compute the velocity [30.37].

Force SensorsForce sensors are used in haptic devices as the operatorinput to an admittance-controlled device, or as a mech-anism for canceling device friction and other undesirabledynamic properties in an impedance-controlled device.Force sensors are described in Sect. 20.4. When a forcesensor such as a strain gauge or load cell measures theoperator’s applied force, care must be taken to thermallyisolate the sensor, since thermal gradients in the sensorcaused by body heat can affect force readings.

30.2.3 Actuation and Transmission

Haptic devices are differentiated from traditional com-puter input devices by actuators that are controlled toprovide appropriate haptic sensations to the human op-erator. The performance of the haptic device dependsheavily on the actuator properties and the mechan-ical transmission between the actuator and the hapticinteraction point (HIP).

Requirements for HapticsThe primary requirements for actuators and mechanicaltransmission in impedance-type haptic devices are: lowinertia, low friction, low torque ripple, back-driveability,and low backlash. In addition, if the design is such thatthe actuator itself moves as the user’s position changes,a higher power-to-weight ratio is desired. Althoughclosed-loop force control has been used for haptic dis-play in impedance devices, most often the mechanism isdesigned to have sufficiently low friction and inertia sothat open-loop force control is accurate enough.

One common mechanical transmission for haptic de-vices is the capstan drive (Fig. 30.6), which consistsof smooth cables wrapped around pulleys of differingdiameter to provide a gear ratio. A no-slip, high-frictioncontact between the cable and the pulleys is maintainedthrough several wraps of the cable. The capstan driveminimizes friction forces felt by human operator be-

DC motorwith encoder

Thermalinsulator

Capstan drive

Axis ofrotation

Loadcell

Fig. 30.6 This version of the Haptic Paddle [30.38] in-cludes an encoder for position sensing, a single-axis loadcell for force sensing, and a brushed motor with capstantransmission for actuation

cause it prevents translational forces on motor and jointaxes.

Current amplifiers are typically used to create a di-rect relationship between the voltage output by thecomputer via a digital-to-analog (D/A) converter andthe torque output by the motor. The effect of actuatorand amplifier dynamics and D/A resolution on systemstability is typically negligible in comparison to posi-tion sensor resolution and sampling rate for most hapticdevices. Actuator or amplifier saturation can produceundesirable behavior, particularly in multi-degree-of-freedom haptic devices where a single saturated motortorque may change the apparent geometry of virtualobjects. The force vector, and thus the correspondingactuator torques, must be scaled appropriately if anyactuator is saturated.

30.2.4 An Example Device

As an illustrative example, we will provide detailed de-sign information for a simple one-degree-of-freedomhaptic device known as the Haptic Paddle [30.38]. Thissection is meant to provide a concrete description ofthe types of components that are used in kinesthetichaptic devices, and the device can also be constructedfollowing the instructions provided by John HopkinsUniversity [30.39]. Many widely available haptic de-vices share the common working principles of thisdevice and differ chiefly in kinematic details arising froma greater number of degrees of freedom.

PartD

30.2

capstan drive

Phantom Premium, SensAble Technologies


a version of the haptic paddle

direct drivemotors attached directly to link(s)

Hayward (McGill)


lecture 3: kinesthetic haptic devices: sensors and actuators · lecture 3: kinesthetic haptic ......

Documents