A Low-cost Arm Based Motion Restriction Haptics for Virtual Reality

A wearable flexible exoskeleton that makes virtual objects feel solid.

๐Ÿฅฝ Product:Haptic hardware + VR game, from scratch
๐Ÿ‘ค Team:Individual work, personal project
๐Ÿ’ผ Role:Research, industrial design, prototyping, testing
โฑ๏ธ Time:One winter vacation
Paper published in IEEE-VR 2021, Lisbon Hero shot: device worn on right arm, front view, arm raised

What is unique about this product?

๐Ÿ’ฐ Total build cost:
โ‚น900 / $12
โš–๏ธ Total weight:
<200 g
๐Ÿ›  Design:
DIY

Cost is excluding the VR headset. Product can be built with off the shelf parts.

Quick context

โ“ Why

  • Virtual objects feel like nothing when you hit them
  • Rich haptics stay locked inside labs

๐Ÿ” What

  • Arm-based flexible exoskeleton built from cable ties
  • Provides motion restriction on VR contact

๐Ÿ‘ฅ Whom

  • VR players & enthusiasts
  • Haptics researchers
  • DIY / maker community

๐Ÿ›  How

  • Cable ties that lock one way, actuated by servos
  • Off-the-shelf parts, 3D-printed casing
  • Driven live by the game engine

Product Brief

This is a low-cost device I designed that provides motion restriction haptics, built under 900 INR (12 US Dollar).

Virtual reality requires high-level haptics to achieve a feeling of realism. The device is a wearable, arm-based flexible exoskeleton made using cable ties, which provides the required motion restriction based on VR interaction.

My process

Identified problems
State of art analysis
Ideate solutions
Design and prototype
Interactions testing
Creating use cases

Problem Brief

โœ‹1. Nothing stops your hand

When a player hits an object in VR it's hard to feel the object restricting their motion, since no obstacle exists in the real world in that exact space. This kills the realism.

๐Ÿ“ฆ2. No sense of presence

There is no motion restriction to convey the presence of virtual objects or obstacles โ€” essential, because interactions in VR are rich and direct.

๐Ÿ”ฌ3. Good haptics stay in labs

Rich and sophisticated haptic devices are primarily associated with labs and are not easily commercialised.

When interacting with rigid objects in a virtual reality environment, the hand passes straight through โ€” the body gets no signal that contact happened.

Diagram: hand stopping at a surface, in-game vs real life
Diagram: punching bag obstacle, in-game vs real life

State of Art

Four existing approaches to arm haptics in VR, and where each one falls short.

A new force-feedback arm exoskeleton for haptic interaction in VR

Force-feedback exoskeleton photo
Pain points
  • Heavy machinery
  • Only available in labs
  • Cannot be commercialised

Wireality: complex tangible geometries in VR with worn multi-string haptics

Wireality multi-string haptic device photo
Pain points
  • Only restrictive to the wrist
  • Doesn't provide arm restriction
  • Hindrance in movements

Frozen Suit: toward a changeable stiffness suit and its application for haptic games

Frozen Suit gameplay + real-world photo
Pain points
  • Motion is very restrictive and painful
  • Doesn't provide enough strength
  • Doesn't support large systems
  • Very hard to set up

Elastic-Arm: human-scale passive haptic feedback for virtual environments

Elastic-Arm device photo
Pain points
  • Suits larger physical spaces like CAVE systems
  • Unidirectional โ€” always toward the shoulder
  • Feedback limited to the elastic cable's stiffness
  • Very little control

Design goals

Pulling the state-of-art gaps together with the problem statement gave six goals to design against.

Ideation

Mapping the problem space across three angles before committing to a mechanism.

Ideation Senses Implementation Body parameters Pressure Feeling of hit Proprioception Vibrations Air drag Feedbacks Comfort Torsion spring Fluid pressure lock Solenoids Slotted cables Assembly Ratchet lock String based Weight Studying arms Studying joints Inertia Motion constraints Tensegrity Stretching directions Recoil force while stopping

Considering human factors

Analysing how far each joint actually travels, so the exoskeleton restricts without fighting the body.

Possible shoulder movements

Side view: shoulder movement range

Forearm & wrist

Forearm movement range

Possible forearm movement

Wrist movement, top & side view

Possible wrist movements

*Grey area indicates the movement space

Two mechanisms considered

Ideation using cable ties

Cable tie locking mechanism, unlocked / locked

Using cable ties which lock in only one direction and move freely in the other. Making the head controllable creates the opportunity to lock and unlock by program.

Ideation using torsion spring

Torsion spring auto-retreat mechanism

The problem in the previous idea โ€” the free-hanging cable โ€” can be hidden inside the slot itself using a torsion spring, which facilitates auto-retreat of the overhanging cable.

Design & Implementation

After studying human arm motion and constraints, only the elbow and shoulder joints โ€” which control the whole limb โ€” are considered. The arm is connected with several links, each made of cable ties. One end is fixed on the hinge joint, the other attached with rubber bands for a smooth retreat. The cable moves freely through a slot in the casing, which houses the servo motors. An inverse-kinematics rig with the controller as target achieves synchronous arm motion in VR.

Wrist link arrangement, two hand poses

Arrangement for restriction of wrist movement

Shoulder link arrangement, front

Complete arrangement

Full exoskeleton arrangement, torso + arm
Device worn, front, holding controller

Materials

Selecting materials so they are low cost and easily available โ€” which also makes the DIY model possible.

Nylon cable tie

Nylon cable ties

12 in, 5 pc ยท โ‚น10 ($0.14)

Restraining arm motion requires a large amount of tensile strength, so cable ties were preferred โ€” they lock in only one direction.

Steel cable tie

Steel cable ties

12 in, 5 pc ยท โ‚น20 ($0.28)

Nylon cable ties are too flexible for operation, so stainless steel cable ties were adhered to them to make them rigid enough.

Velcro band

Velcro bands

1 metre ยท โ‚น30 ($0.41)

Velcro forms the bands the locking unit sits on, and holds the device firmly on the user's arm.

Rubber band

Rubber bands

โ‚น5 ($0.069)

Connected to the rear end of the cable to provide a smooth retreat of the cable back inside the slot.

Micro servo SG90

Micro servos SG90

โ‚น90 each ยท 5 units โ‚น450 ($6.21)

Actuates the cable ties. Under 20 g (9 g motor + 8 g cable ties), 0.1s/60ยฐ speed, and only 15ยฐ of travel โ€” so latency is negligible.

Arduino Uno WiFi board

Arduino Uno WiFi

โ‚น350 ($4.83)

Programs the servo motors to fire based on the signals coming from the game engine.

Casing CAD render

Casing (3D printed)

โ‚น100 ($1.38)

Designed in SolidWorks and 3D printed. Holds the motor, the arm, the cable-tie head and the cable ties in the slot.

Complete prototype

Full shoulder-to-wrist harness, unworn, laid flat

This is how the wearable device looks. You just strap the Velcro around your arm and you're ready to go.

It consists of 5 units of cable mechanism โ€” 3 for the shoulder and 2 for the elbow joint. The wrist was not included in this prototype.

Working principle

Use case 1

When bringing the forearm down and being stopped by an obstacle in VR space.

Forearm stopped by obstacle, locking mechanism callouts

Use case 2

Forearm moving toward the shoulder and being stopped by an obstacle in VR space.

Forearm raised toward shoulder, stopped by string tension

Interior working mechanism

Servo unlocked, top view + cross section

Unlock state

The cable-tie head has not been engaged, so the cable ties can move freely in both directions.

Servo locked, top view + cross section

Lock state

The motor detects the signal and exerts a small force on the cable-tie head to engage it on the rack, which completely restricts motion in one direction.

Working mechanism

The same drumming action, with and without the device on.

Drumming without device (video / gif)

Without device

The arm stretches fully with no constraints, even after the stick hits the drum.

Drumming with device (video / gif)

With device

The cable-tie link on the biceps locks, so arm motion is restricted the moment the stick hits the drum.

Interactions & Evaluation

To test the device, 5 types of interaction were built around motion-restriction haptics โ€” the common patterns you find in VR games: holding an object, and interacting directly with the hand.

Drumming
Interactions
With the prototype
Without the prototype
A ยท Hitting objects

Drumming action

Demonstrates collisions with a tool. Uses forearm and elbow movement.

Drumming, with prototype
Drumming, without prototype
B ยท Lifting objects

Lifting a sword

Demonstrates the sense of inertia and gravity of the object.

Lifting sword, with prototype
Lifting sword, without prototype
C ยท Motion of held objects

Short duel with the sword

The player defeats an enemy wielding sword and shield โ€” demonstrates shoulder movement.

Sword duel, with prototype
Sword duel, without prototype
D ยท Hand-based interaction

Pressing a button inside a box

The box has obstacles on four sides โ€” demonstrates touch and lateral motion restriction.

Button in box, with prototype
Button in box, without prototype
E ยท Inertial recoil

Punching a bag

Demonstrates sudden recoil haptic feedback.

Punching bag, with prototype
Punching bag, without prototype

Product stats

*This prototype is for single arm use case, Idea can be extened to whole body use case

  • Cable-tie links actuated by micro servos5 units
  • Joints covered โ€” shoulder & elbow3 + 2
  • Servo travel to engage the lock15ยฐ
  • Actuation latency at 0.1s/60ยฐNegligible
  • Interactions built & tested in VR5 types

Designed for DIY, Every part is off-the-shelf, so anyone can build one

  • Casing designed in SolidWorks3D printed
  • No soldering โ€” Arduino + servo wiring onlyBeginner-friendly
  • Total unique parts to source7
  • Assembly time per unit~15 min

Future scope & conclusion

Future scope

The prototype does not include any finger-based motion restriction โ€” it predominantly focuses on forearm and shoulder movement.

The concept and prototype were tested on a limited set of users so far. More tests are planned with people of different physiques and proportions.

Conclusion

The prototype's whole build is very cheap โ€” made under 900 INR (12 US Dollar).

Given the build and cost, it can easily become a commercialised product, or a DIY model once the 3D print files are shared.

Touch and Explore: a VR game built on haptic-driven gameplay

Experience a world without light, and navigate using touch, motion restriction and proprioception.

Paper published in ACM Interactive Surfaces and Spaces 2022, Poland Dark VR scene, silhouette with light-wave touch cues
๐Ÿ”— Context:Use case of VR-Haptics project
๐ŸŽฎ Type:Game design + haptic interaction
๐Ÿ›  Built in:Unity + XR kit, Oculus Rift
๐Ÿ’ก Idea:Zero visuals, pure haptic navigation
๐Ÿ‘ค Team:Individual work, personal project

My process

Game ideation
Refining game mechanics
Prototype
Testing gameplay in small loop
Final design

Touch & Explore game design

Touch and Explore is a VR game driven by haptic navigation โ€” the sense of touch, motion restriction, and proprioception.

Core game loop

01 Player entersdark space,with no light 02 Touches theenvironment toexplore thetopology 03 Gains visualcues of theenvironment 04 Completes theobjective forthat level

The logic behind it

  • At the start the player is in a dark environment and can see only their hands and the instructions.
  • Explore the game world only through touch and motion restriction, using the revealed topology to navigate.
  • Visuals are reduced to almost nothing, so immersion comes purely from haptics โ€” the whole game takes place in dark space.
  • The player never knows what is coming, so every second is a curiosity trigger โ€” an important factor in gameplay experience.
Dark scene, glowing hand touching a surface
Player standing inside Touchscape, second camera view

A player standing inside the Touchscape VR game, captured from a second camera. The light waves indicate recorded touches.

Note: the game is completely dark โ€” ambient light has been added here purely to make the environment readable.

Inputs, development & sound

Inputs

  • Since the whole game is haptic driven, the player's interaction radius is reduced to their arm length in the game world.
  • Only when the player discovers what's in their immediate vicinity can they move further and increase the interaction space.
  • Once an obstacle is found, its position has to be memorised โ€” visualised as touch-points, in the form of light waves emerging from the point of contact.
  • Standard VR teleport control becomes available once a space is discovered through touch, plus a click button for options.
Hand touching surface with ripple cue

Touching a surface starts to reveal the object's presence through visual cues.

Input hardware

HMD gear
VR controller
Haptic prototype
Device worn, headset + controller ready to play

Development

  • Designed in the Unity game engine with the XR kit. Oculus Rift used for the VR gear.
  • Every joint in the motion-restriction gear is mapped to the player's joints using an inverse-kinematics rig, so the avatar's arm follows the real arm.
  • If the player collides with or touches any object, the game engine triggers the respective link to lock โ€” restricting motion.

Sound design

  • With visual elements reduced, audio carries the sense of presence in VR.
  • Engaging ambient sound keeps the player entertained in dark space โ€” the way a person enjoys music more with their eyes closed.
  • Touch and texture map to different sounds: hitting a concrete wall gives a hard, high-pitched sound; touching cardboard gives a low, soft one.

Game play

Gameplay frame 1

How a session plays out

  • The player enters a completely dark environment without any light, shutting off their sense of sight.
  • In-game instructions are the only visible cues at the start.
  • The player can move their arms freely โ€” they see a blue transparent hand, giving a sense of proprioception in the Touchscape.
  • As they move, they touch invisible obstacles to reveal the topology, felt through motion restriction.
  • Each obstacle hit is recorded in memory and visualised as light waves โ€” visual cues from contact alone.
  • Players then move around and explore using touch and motion restriction to finish the mission, such as escaping the room.
  • As the game proceeds, new elements are introduced โ€” weapons, enemies, and so on.

Level designs planned for the future

  • Levels with limited visual memory
  • Limiting the visual cue of touch
  • Levels with a time duration on the visual cue
  • Maze-like levels
  • Escape-room levels

Limitations & future scope

Limitations

Players initially felt disturbed starting a game in a dark environment without any light โ€” they slowly adapted to it.

Visual cues are quite intense right now; the intensity can be programmed down.

Exposing players to a completely lightless environment for a long time might have drawbacks. Prolonged research is planned to identify any problems.

Currently the game deals only with arm-driven haptics. Exploring the concept with full-body haptics is planned.

The concept and prototype were tested on two players, so more tests are planned.

Future scope

This is a novel game concept rather than a full-fledged game โ€” it deals with haptic-driven gameplay in a VR setting.

The game is highly interactive with the space around the player, and playable only with motion-restriction haptics โ€” a prototype made specially for it, at a time when similar haptic inventions are blooming.

It could create an opportunity to empathise with visually challenged people, since the gameplay is somewhat similar to what they experience.

Touch and Explore can be a stepping stone for exploring immersion in games using only haptic senses, without visuals.

Combining the concept with spatial audio and echolocation could create even deeper immersion โ€” arguably a new genre.

Reflections & learnings

๐Ÿงฐ Cheap parts can carry a real idea

Cable ties, rubber bands and โ‚น90 servos delivered genuine motion restriction. Constraining the budget forced clarity about which mechanism actually mattered.

๐Ÿ“ Studying the body first saved iterations

Mapping shoulder, forearm and wrist ranges before designing links meant the exoskeleton restricted motion without fighting natural movement.

๐ŸŽฏ Building a use case proved the hardware

A haptic device is hard to argue for in the abstract. Designing Touch and Explore around it turned a prototype into a demonstrable experience.

๐Ÿ”„ Testing across interaction types exposed gaps

Five distinct interactions โ€” hitting, lifting, wielding, pressing, recoiling โ€” surfaced limits a single demo would have hidden entirely.

๐Ÿค Hardware and software have to be designed together

The inverse-kinematics rig, the servo travel and the game's collision events all had to agree. Treating them as one system, rather than a device plus a game, is what made the restriction feel instant.

Thankyou