Position Sensing(Take class notes)Type of Position Sensing• There are five fundamental types of position sensing elements that are used in robots- Potentiometers- Linear Variable Differential Transformer (LVDT)- Resolvers- Absolute Optical Encoder- Incremental Optical EncoderPotentiometers• There are two types – linear and rotary. The latter is more often• Use mechanically variable resistors• Material: wire-wound; thin films of carbon, metal, or other conductive materials• Linear 0.1% accuracy• Rotary – cannot turn beyond 360 – in fact it is often less: space occupied by contacts;- Alternate – 5 to 10 turn helix or with gearing which converts linear motion to rotary motion- Provide absolute position when power is off- Low cost; $5 – 25- Life is less than non-contacting sensorLVDT• Linear Variable Differential Transformer (LVDT)- Made up of 3 coils- AC supplied, and DC out- DC – linear function of displacement of core (coupling changes). Note: no voltage when core is symmetrically placed- Good: no contacting parts; bad: more expensive- Keep absolute position even with power offResolver• AC input, amplitude of sine and cosine outputs vary with the angle• Complex electronics – including analog and digital (resolve digital)- Special IC circuits are available for the purpose- Analog Devices: Variable Resolver-to-Digital converters: AD2S81A/AD2S82A • 12-bit resolution (1 in 4096)• No problem with power – off• 0.1% accuracy• Brushless resolvers eliminates arc and wear• 2 fixed coils – 90oapart (stator)- Rotor coil produces: - The voltages of the stator coils: = and =Absolute Optical Encoders• Multiple concentric rings of binary code are photographed/etched on the code wheel; radical line (angle) – absolute value of the position• No problem with power-off; very accurate• 10-20 rings - 210to 220resolutions• Photolithography – expensive - Start at $300• Need N light sources and N detectors (N is the number of rings)• Gray code could be used – leads to 1 bit change at a time – more accurate- Consider moving from 0111 1000- Only increased by 1, but all the bits changed- Using Gray code we can have 0111 0101Only one bit changeA2A3A0A1MotionIncremental Shaft Encoders• Installed on the motor shaft• Simpler and less expensive• Two photo cells – sense direction & individual steps• Two-track structure – with slots- Each slot width = 2 sensor (photo cell) widthIncremental Shaft Encoders (2)• One-track structure (2 photo cells)- sensors = ½ slot width; displaced usually by 1 slot.• In general small gear ratio between joint and encoder high positional accuracy for reasonable encoder sizeV(A) V(B)Visual Sensors• Visual position sensors have been used in robotics for a long time• Most of the visual sensors such as a cameras are used for navigating a mobile robot – low precision in a few centimeters• Machine vision is a term saved for robots and factory automation – high precision up to fractions of a millimeter- provide imaging-based automatic inspection- provide robot guidance in industry – vision based feedback controlRobotic systemPID controlVision-basedControlCameraInputWhat about Velocity Sensing?• Several approaches- Tachometer – a small DC generator- Use time between encoder pulses- Back EMF of driving motor- Derivative of position
View Full Document