Unit 7: End Effectors and Sensors

I. Orientation: Robotic Manipulation and Perception

Robotic arms perform useful work through two complementary systems: an end effector physically interacts with an object, while a sensor measures the robot, object, or surrounding environment. In the Orangewood Robotic Arm laboratory, these systems are studied by attaching different tools, commanding arm motion, observing grasping or manipulation, and interpreting sensor feedback. The governing principle is the sense–plan–act cycle: the robot senses conditions, selects or follows a motion plan, and acts through its joints and end effector.

• End effector: The device mounted at the robot’s wrist or tool flange; examples include grippers, suction cups, magnetic tools, and pens.
• Sensor: A device that converts a physical quantity such as distance, force, light, angle, or temperature into an electrical signal.
• Robot coordinates: Position is commonly described by Cartesian coordinates ((x,y,z)), while orientation may be represented by roll, pitch, and yaw.
• Tool center point (TCP): The reference point used by the controller to describe the working position of the attached tool.
• Feedback: A measured output is returned to the controller so that motion or gripping can be corrected.
• Payload: The total load carried by the arm, including the object and the end effector; it must remain within the arm’s rated capacity.
• Repeatability and accuracy: Repeatability describes returning to nearly the same location; accuracy describes closeness to the commanded or true location.
• Safety convention: Motion should be tested at low speed, with the workspace clear, emergency stopping available, and tool attachments securely mounted.

II. End Effectors — Tools for Physical Interaction

A. Study and usage of different types of end effectors and sensors using Orangewood Robotic Arm

An end effector is selected according to the object’s shape, material, weight, surface, and required operation. The Orangewood Robotic Arm provides the motion, while the mounted tool determines how that motion becomes grasping, lifting, placing, writing, or inspection.

• Mechanical gripper: Uses fingers or jaws to hold an object by friction or enclosure.
  • Parallel-jaw gripper: Its fingers move toward each other along approximately parallel paths and is suitable for blocks, cylinders, and regular components.
  • Angular gripper: Its fingers rotate about pivots; it can be compact but may produce changing contact geometry during closure.
  • Three-finger gripper: Provides three contact points and can center cylindrical or irregular objects more effectively than a two-finger tool.
• Gripping principle: A friction grip must generate enough normal force to prevent slipping. A simplified condition is:

  F_friction = μN

Here, (F_{\text{friction}}) is the available friction force in newtons, (\mu) is the coefficient of friction, and (N) is the normal gripping force. With two identical contact surfaces, the approximate holding force is (2\mu N), ignoring acceleration and tool geometry.

• Vacuum or suction gripper: Uses a pressure difference between the suction cup and atmosphere to hold smooth, nonporous surfaces such as glass, plastic, or finished metal.
  • Holding force: Approximately (F = \Delta P A), where (F) is force in newtons, (\Delta P) is pressure difference in pascals, and (A) is effective cup area in square metres.
  • Limitation: Porous, rough, dusty, or highly curved surfaces may leak and reduce the gripping force.
• Magnetic end effector: Uses a permanent magnet or electromagnet to lift ferromagnetic materials such as iron or steel.
  • Advantage: It can pick up flat metal parts without surrounding fingers.
  • Limitation: It does not work reliably with aluminium, copper, wood, plastic, or nonmagnetic stainless steel.
• Adhesive or electrostatic tool: Uses adhesive material or electrostatic attraction to handle thin, lightweight, or delicate objects.
  • Limitation: Adhesive contamination, surface cleanliness, humidity, and object weight affect performance.
• Tool or process end effector: A pen, screwdriver, welding tool, cutter, dispenser, or polishing head performs a process rather than simply holding an object.
  • Example: A pen-mounted tool can trace a programmed path, but the arm must control position and, where necessary, contact pressure.
• Mounting and calibration: The tool must be firmly fixed to the wrist, and its TCP must be defined relative to the flange.
  • Effect of incorrect TCP: If the controller assumes the tool tip is 50 mm from the wrist when it is actually 70 mm away, the commanded path can miss the object even though the joint angles are correct.
• Operating procedure: A typical laboratory sequence is to mount the tool, check its alignment, set or verify the TCP, select a low speed, approach the object, close or activate the tool, lift vertically, move to the destination, release, and return.
• Selection criteria: The correct end effector depends on object mass, dimensions, surface texture, fragility, required grip, cycle time, power supply, and the possibility of tool collision.
• Safety limitation: A gripper can pinch fingers, a magnetic tool can attract loose metal, a suction system can release suddenly, and a process tool can cut, heat, or damage nearby objects.

B. Applications and limitations

End effectors determine the practical capability of the Orangewood Robotic Arm, but they cannot compensate for poor positioning, inadequate payload, or unsuitable object properties.

• Pick-and-place: A gripper is appropriate when an object must be lifted, transported, and placed at a defined location.
• Sorting: Different tools can separate objects by material or shape; a magnetic tool, for example, can distinguish ferromagnetic parts from plastic parts.
• Assembly: A narrow finger tool can insert a component, but excessive force or positional error may damage the component.
• Demonstration and teaching: Interchangeable tools show how the same arm can perform different tasks without changing its joint structure.
• Failure sources: Slippage, inadequate suction, tool flexure, inaccurate TCP calibration, excessive acceleration, and collision are common causes of unsuccessful operation.
• Performance measure: A useful evaluation records successful picks, dropped objects, cycle time, placement error, and damage to the object.

III. Sensors — Measurement and Feedback

A. Study and usage of different types of end effectors and sensors using Orangewood Robotic Arm

Sensors allow the Orangewood Robotic Arm to obtain information that cannot be supplied by a motion command alone. They may be built into the arm, attached near the tool, or placed in the workspace. Their readings are used for position control, object detection, collision prevention, grasp verification, and process monitoring.

• Position or encoder sensor: Measures joint angle or motor-shaft rotation.
  • Incremental encoder: Produces pulses as a joint moves; position is calculated by counting pulses from a reference.
  • Absolute encoder: Provides a coded position value that remains available after power is restored.
  • Joint relation: For a rotary joint, angular position may be estimated as (\theta = 2\pi n/N), where (\theta) is angle in radians, (n) is counted encoder units, and (N) is encoder units per revolution.
• Proximity sensor: Detects an object without physical contact.
  • Inductive proximity sensor: Detects metallic objects using a changing electromagnetic field.
  • Capacitive proximity sensor: Detects changes in capacitance and may respond to plastic, wood, liquids, or other materials, although humidity and material properties affect it.
• Ultrasonic distance sensor: Emits an ultrasonic pulse and measures the returning echo. Distance is estimated by:

  d = vt / 2

Here, (d) is distance in metres, (v) is the speed of sound in air, approximately (343\ \text{m/s}) near room temperature, and (t) is the round-trip time in seconds. Division by two is required because the pulse travels to the object and back.

• Infrared or optical sensor: Uses emitted or ambient light to detect an object, estimate distance, or identify contrast.
  • Reflective sensor: Measures returned light; black, shiny, transparent, or angled surfaces may produce unreliable readings.
  • Break-beam sensor: Detects an object when it interrupts a light path between an emitter and receiver.
• Vision sensor or camera: Produces an image from which shape, colour, position, orientation, or presence can be estimated.
  • Coordinate conversion: A camera pixel ((u,v)) must be calibrated or transformed into robot coordinates before the arm can accurately pick the object.
  • Limitation: Lighting, shadows, occlusion, lens distortion, and similar object colours affect recognition.
• Force or torque sensor: Measures contact force or wrist torque during gripping, insertion, or collision detection.
  • Force feedback: The controller can stop or reduce motion when measured force exceeds a safe threshold.
  • Limitation: Sensor readings may include tool weight, gravity, vibration, and mounting errors.
• Tactile sensor: Detects contact, pressure distribution, or finger touch. It can confirm that a part is inside the gripper even when a camera cannot see the contact.
• Temperature sensor: Measures temperature in a process or component; it is relevant when the tool handles heated parts or performs thermal operations.
• Digital and analog outputs: A digital sensor typically gives two states, such as 0 for no detection and 1 for detection. An analog sensor provides a continuously varying voltage or numerical value related to the measured quantity.
• Calibration: A sensor reading must be compared with a known condition.
  • Offset calibration: If a distance sensor reads 12 mm when no object is present at the reference position, that offset must be removed.
  • Scale calibration: Two known distances can be used to estimate the conversion between output voltage and actual distance.
• Filtering: Noise can be reduced using an average of recent readings:

  filtered = (r1 + r2 + ... + rk) / k

Here, (r_1) through (r_k) are the latest sensor readings and (k) is the number of samples. A larger (k) smooths noise but increases response delay.

• Sensor integration: A practical pick operation may use encoders for arm position, a proximity sensor for object detection, and a tactile or force sensor to confirm a successful grasp.

B. Applications and limitations

Sensor usage changes the arm from a purely pre-programmed mechanism into a system capable of responding to actual conditions in the workspace.

• Object localization: A camera or distance sensor can identify where an object is located before the arm approaches it.
• Approach control: Proximity feedback can slow the arm near an object, reducing collision risk and positional error.
• Grasp confirmation: Tactile, force, or motor-current feedback can indicate whether the object has been captured rather than merely surrounded by the fingers.
• Collision detection: An unexpected force or torque can trigger a stop, but the threshold must distinguish a real collision from normal acceleration or tool weight.
• Sorting and inspection: Vision, inductive, and colour sensors can classify objects before a suitable end effector performs the next action.
• Sensor limitations: Every sensor has a range, resolution, response time, field of view, and susceptibility to environmental interference.
• Practical reliability: Sensor data should be checked for impossible values, sudden jumps, disconnected signals, and readings outside the calibrated range.
• End-effector interaction: The sensor and tool must be considered together: a suction cup may obscure a camera, a metal gripper may interfere with an inductive sensor, and flexible fingers may change force readings.
• Laboratory evaluation: The experiment should compare commanded position with measured position, record detection success, and identify how lighting, object material, distance, and tool alignment affect results.