Blinds' Stick, With Laser, Tactile

[Enthalpy]

Hi dear readers!

To improve blinds' stick, lasers and previously acoustic sensors seem an obvious choice, to detect and range objects and obstacles sooner.
http://www.lac.u-psud.fr/teletact/publications/report_training.htm (among many)
Though, the varied designs have not replaced the white stick.

One reason could be that the acoustic signal that represents the distance competes with useful surrounding sounds and is difficult to use to make a mental image of the surroundings. It would also have to represent very fast variations: if the user scans 5m width in 1/4s, a 0.1m pole lasts for 5ms only, which is difficult to perceive through a pitch change - such a fast sweep is needed if scanning through height as well. Some designs add a tactile actuator, but from what I've read, only as a warning vibration.

I propose (is that new?) to represent the distance by the movement of a part that the user feels with a finger.

• People are already trained to build mental images from tactile feelings, blinds even more so.
• Voice coil motors and piezoelectric actuators can be very fast.
• Tiny movements are easily perceived, quick ones as well.
• The stick just indicates in real time the distance in front of it.

 

 

Ergonomics is very hard to predict... Much must be determined experimentally.

• Whether this idea is useful at all...
• Whether the part's movement shall be parallel to the distance and in the same direction.
• What amplitude, reaction time, position relative to a fixed reference.

2mm move in 5ms is feasible but would draw much on the battery; piezo actuators look better for that. Anyway, our tactile sense doesn't react so fast, so I expect the user to detail more slowly a small object once detected, and the actuator needs not be so swift - some reaction as fast as the tactile sense, if not at full amplitude, should suffice. We feel and act much faster than 1/10s, as musicians know, but 1/100s is a wide lower bound. Signal tweaking, like nearer narrow objects being represented a bit longer, looks interesting.

A big maximum range would improve over the white stick, say to orient oneself in a road crossing, but short-range accuracy is necessary. The distance measure can be transformed nonlinearly, for instance through a logarithmic compression, for better perception. Several moving actuators, possibly for different fingers, can also work at different ranges.

The light shall be broad enough to prevent harm, but diverge rather little. Visible light must be better accepted by surrounding people. Such a "stick" would better fit in the palm.

Marc Schaefer, aka Enthalpy

[John Cuthber]

Next time I see a blind person with a stick I will ask what the battery life is like.

[Enthalpy]

No electronic apparatus can challenge a white stick on battery life, I approve that...

 

But a battery that lasts for instance 16 hours and recharges overnight makes a usable apparatus, and the available power exceeds the expectation.

 

From Conrad's catalogue, Li-FePO₄ accumulators used by remote control hobbyists store about 370kJ/kg, so:

• 500g carried at the vaist provide 3.2W
• 50g in the hand-held box provide permanent 320mW

which is more than enough for a laser, electronics and a well thought actuator. For instance the powerful laser diode in a CD burner makes 20mW light with excellent efficiency - and we need only short light pulses.

 

The power consumption is easier than I feared. I'll give a thought at the actuator.

[Enthalpy]

This voice coil motor can move the button that represents the distance. To fit in a narrow long box, it splits the coil in two and the flux path in four, achieving 1.0T at the coils' mean D=20mm.

 

A permanent force of 1N takes only 21mW losses, and accelerates the ~32g mobile mass by 31m/s². If limiting the acceleration force to 1N and the speed to 0.15m/s, the full 2mm stroke takes 18ms.

The full speed and force take (or restore) 150mW, exceeding the losses, which suggests regenerative braking. The voice coil motor permits it; other actuator types need less power or none to keep a constant force, but those who store much reactive energy may complicate regenerative braking. A spring balancing the finger's load would also reduce the permanent force.

The small power suggests to recharge the battery by a handle or equivalent.

An integrated class D audio amplifier can drive the voice coil motor; it needs output Mos with low resistance to save power when braking or when the button is immobile. I'd make a full servo loop on the button's position, but with force limits clearly felt by the user.

If wasting more power, the voice coil motor can be smaller. For instance, one half of the above design would consume 2*21mW for permanent 1N. The loudspeaker's design, wide and short, would require a playless angled lever, or a different orientation of the box.

Marc Schaefer, aka Enthalpy

[Enthalpy]

Some people suggested me to use ultrasound instead of a laser. This would have drawbacks:

• Sound takes 0.1s to reflect from 17m distance, so the user could not scan his environment quickly enough with his own movements;
• The angular resolution is bad. At 340kHz (already much for air propagation), a 50mm source, reflector or lens still makes the main lobe 2*0.24m wide at 10m distance.
• Acoustics is plagued by parasitic reflections. Some signal processing methods can alleviate them, nothing easy nor perfect.

so light is generally better. If using sound, say in environments of bad visibility (fire brigade), scanning can be fast enough:

• Send a ping in all directions, listen to the echoes from all directions simultaneously, as Sonar do it for the same reason;
• Reconstruct the echoes from individual directions by signal processing, called "aperture synthesis", as Sonar and Radar do it;
• Store the information about the environment, give it to the user one at a time, just in the direction he designates;
• For that, the user may have a gyrometer or an other orientation sensor, for instance at his hand, and for example a tactile actuator in a glove.

 

Marc Schaefer, aka Enthalpy

 

 

[Enthalpy]

Could the apparatus help divers? Visibility is often poor where they need to operate, and sound propagates well underwater. Divers may (or not) prefer a tactile information to a visual display.

For a limited range like 20m, high frequencies are possible. Echoes are a lesser worry in water, but should still be processed away.

Marc Schaefer, aka Enthalpy

[Enthalpy]

The following actuators with mobile coil are smaller and lighter. The battery is but heavier as the actuators draw more power for a constant force.

The cylindric actuator, left side of the sketch, has a single D=10mm L=50mm magnet to create all 4.5 poles: easier to assemble. CyMagnetics' N38 material (915kA/m and µ=1.06) creates 0.55T at the coils' 13.5mm mean diameter - the induction dilutes from the 4mm long poles, but the 8mm long coils catch it. Iron poles were useless for this induction.

Constant 1N takes 100mW. The actuator fits in D=18mm, weighs 70g, with 23g for the mobile part.

----------

A flexible printed circuit makes the coils set. 50µm patterned copper over 25µm of polyimide or polyester film use well the volume. The film is rolled to make the cylindre without keeping a coil former; adhesive tape, liquid glue, or pre-molten or re-molten glue, makes a stiff solid part. 33 turns would bring 2.5mm thickness (with 0.5+0.5mm clearance) and a few tracks connected externally adapt to the supply voltage, so this part of the sketch is not to scale. A flexible circuit is also a reliable way to connect the moving part.

If the tracks keep their respective positions over the pole sequence, they can be powered individually, say to use current only in the tracks that receive a strong induction at a given coil displacement. The stroke can even exceed a pole spacing if the current is properly inverted - better than a voice coil motor, this is now a multiphase linear motor, smaller and of more flexible design.

The tracks must cross an other to keep the relative positions. A double layer flexible circuit does it. Its lower and upper copper can be already coated or get a separate insulation layer: an extra film, the adhesive tape, or a fabric soaked with liquid, pre-molten or re-molten glue.

----------

A coil or coils set made of flexible printed circuit has more uses, for instance at loudspeakers, other voice coil motors and generators, hard disk drives' arm actuator, and also at rotative actuators, motors and generators, including bell-shaped rotors, other low-inertia motors, and fast-rotating motors.

 

On the right side of the sketch, the flat actuator resembles a hard disk drive's arm motor, but has three pole pairs for convenient shape, and the coils translate, which seems better than rotate here.

The flat design is more adaptable. The example uses 4mm thick N38 material, where 2*3 magnets give their 25mm width and cumulate three 10mm lengths to span three poles of 6mm, 12mm and 6mm length separated by 3mm leak stoppers. This time the coil bars are 4mm wide, narrower than the magnetic poles. The lower half-stator mirrors the upper, and they achieve together 0.8T, so the actuator uses 71mW for 1N, weighs 73g with 12g for the mobile part, and is thin and short.

Copper wire could make the coils set, for instance coated with glue, or included in a thermoplastic injection like in hard disk drives... But a thin printed circuit makes the solid coils pair easily and fits small production series. Stacks of spiral-patterned circuits (sketch simplified) would be hard to connect; a double-sided circuit instead can be folded to make the connected stack - or better, winded flat to make the stack, possibly keeping some bending radius at the edges outside the stator. The tracks can be locally wider. The refinements with tracks powered individually apply here as well.

Marc Schaefer, aka Enthalpy