CA2098153C - Fiber optic curvature and displacement sensor - Google Patents

Abstract

A curvature or bending and displacement sensor is composed of a fiber optic or light wave guide, for attachment to a member or members being curved or displaced. Light is injected at one end and detected at the other end. Curvature of the fibre results in light loss through an emission surface or surfaces, sometimes in conjunction with a superimposed curvature in a plane other than that of the curvature to be measured, this loss being detected. The loss of light detection is used to produce indication of curvature or displacement. The light emission surfaces extend in a substantially circumferential direction, or in a substantially curved axial direction when in a curved portion of a curved guide. The placement, shape and configuration of the emission surfaces allows adjustment of the linear range of measurement, the overall throughput of light, and the length over which curvature is measured. Two or more light guides can be oriented to given indication of direction of curvature or displacement.

Description

FIBER OPTIC CURVATURE AND DISPLACEMENT SENSOR
FIELD OF THE INVENTION
The present invention relates to fiber optic sensors for measuring curvature (bending) and displacement, with high sensitivity and minimal residual light loss, coupled with ease of manufacture and ability to sense near the distal end of a fiber loop.
BACKGROUND OF THE INVENTION
It is often desirable to use fiber optic sensors in or on materials or structures whose strain, curvature or displacement must be measured. Fibers are ideal for many applications because they can be relatively inert to environmental degradation, are light in weight, are not affected by electromagnetic interference, carry no electrical current, and can be very small and flexible, thus having little or no effect on the structure in which they are I embedded. It is possible to either cement fibers to surfaces or to embed them inside, such as in fiber/epoxy composites, concretes, or plastics.
Many sensors are based on measurements of strain, which is basically an elongation of material. Although it is possible to use multiple strain gauges to infer curvature from strain, it is more desirable in many circumstances to measure curvature directly. Often, it is desirable to mount a sensor near the neutral axis of a beam, where there is no strain associated with curvature of the beam. Often, curvature is the parameter of direct interest, such as when measuring ~098~.5~
deviation from straightness in a pipe or rod. It is also frequently desirable to measure displacement between two structures, which can be inferred from the curvature of a flexible beam or fiber connecting them. Just as strain gauges can be used to infer curvature in some circumstances, curvature sensors can be used to infer strain.
Strain gauges have found wide application in a huge variety of measurement tasks; curvature sensors potentially have just as many applications. The following examples cover only some of the potential applications for fiber optic curvature and displacement sensors; measuring flutter and deflection in aircraft wings and aerospace truss structures;
measuring deflections on cranes and lifting devices; measuring movement of bridges, dams, and buildings due to earthquakes, settlement, or other degradation; measuring sag and deflections of pipes, rods, cables, and beams; measuring the effects of frost heave on roadways and runways; sensing traffic movements and soil settlement; measuring wind forces on masts and towers; sensing parameters of sports equipment including skis, poles, shoes, fishing equipment, swords, bats, clubs, balls and clothing; measuring deflections on marine equipment including masts, spars, cables, hull plates, struts, and booms; measuring curvatures of vanes, wires, poles and other flexible structures or probes to infer fluid or slurry flow, speed, and direction of movement; measuring vibration and sound levels by means of flexing beams, fibers, or diaphragms; measuring pressures by sensing the curvature of diaphragms or tanks; measuring acceleration in general;

~~~~~~J

measuring deceleration and associated forces for the deployment of airbags; measuring the deflection of support structures to infer applied weight, forces, torques, and deflections; forming multi-degree-of-freedom force and torque sensors; forming input devices for computers including joysticks, keyboards, and levers; measuring joint angles and deflections on robots, automatically guided vehicles, automobiles, trucks, tanks, earth moving equipment, loaders, cranes, ships, airplanes, helicopters, and spacecraft;
measuring the deflections of tire treads and other rubber or elastomeric moving parts; measuring door and wheel positions;
measuring pedal, vane, rudder, lift surface, and valve positions; measuring shaft and knob angles, rotations, and posit9.ons; measuring liquid levels by deflection of floats or bladders; measuring alignment of automotive, marine, or industrial equipment; measuring positions and motion of reclining seats, chairs, beds, and medical fixtures;
instrumenting medical tools; instrumenting prosthetic devices;
measuring deflections in the presence of high magnetic fields;
and many others.
Generally, fiber optic sensors that must be exposed to harsh environments or that must be embedded, should be intrinsic sensors, that is, sensors that do not rely on light leaving and then reentering the fiber. Thus, sensors that involve light exiting a fiber, reflecting off a surface, and re-entering are not desirable for many purposes because the surfaces may become contaminated, thus changing the light A wide variety of intrinsic fiber optic sensors has been described, most of them based on interference techniques.
Interference-based sensors, which rely on mechanical changes to the fiber dimensions producing changes in light interference patterns within the fiber, are very sensitive to strain but also to temperature and involve complex and expensive electronic circuitry. Another drawback is that usually lasers or laser diodes must be used as light sources with these sensors, thereby limiting their durability and longevity, and increasing their cost.
Optical fibers transmit light by virtue of total internal reflection. The light is contained in a core of transparent material. Generally, this core is covered with a cladding layer that has a lower index of refraction than the ~ core. Because the index of the cladding is lower than that of the core, rays within a certain range of angles of incidence with the core/cladding boundary will be refracted back into the core upon striking the cladding. If the fiber is bent in a curve, small amounts of light are lost due to changes in the angle of incidence at the curved core/cladding boundary. If the curvature becomes substantial, significant amounts of light may be lost. Fibers with a discrete core/cladding boundary are called step index fibers. Other fibers called graded index fibers do not have a distinct boundary between core and cladding, but exhibit a continuous decrease in index of refraction toward the outer circumference of the fiber.
For simplicity, this description will use terminology consistent with step index fibers, but graded index fibers may be similarly treated, as may other light guides including guides of non-circular cross section or guides with gas or liquid surrounds instead of conventional solid cladding. It is also possible to use metal-covered fibers.
"Microbending" sensors are designed to take advantage of this loss mechanism. They generally involve a mechanical structure such as a serrated plate that presses on the fiber, producing a series of substantial local curvatures (bends). The loss of light is used as a signal to indicate displacement of the mechanical structure. Microbending sensors generally do not have a linear loss of light energy in response to changes in curvature, and are otherwise undesirable because of the necessity for a mechanical structure, and the strain which it imposes on the cladding and core of the fiber during deflection. If fibers are used without a mechanical structure to translate displacement into large local curvature, then the light loss due to bending is sufficiently large to be of practical use only when the bending is large. For small bends, such unenhanced microbending sensors are not.useful because inadvertent bending of the fiber optic leads carrying light to and from the sensor portion of the fiber will produce changes in light loss that are indistinguishable from those produced by bends of the sensor portion. For these reasons, microbending sensors are generally not used in embedded applications, and rarely are used for measurements of curvature.
It is possible to treat optical fibers so that the amount of light travelling through the core changes more than 2~~~~53 usual with changes in curvature. Methods generally involve modification of the cladding so that it loses more light than usual over a short length. When straight, more light than usual is lost over the treated zone. When bent, additional amounts of light are lost due to the greater interaction of the treated sides with the light travelling through the fiber.
Methods of treatment include abrasion, etching, heat treatment, embossing, and scraping of the cladding. Such treatment can produce a loss of light that is linear with curvature over a wide range, and which is much greater, by orders of magnitude, than the loss produced by microbending or by inadvertent bending of the leads carrying light to and from the sensitized zone.
A drawback of the above method of treatment is that 15' loss is introduced even for a straight fiber, and the modification of the cladding can weaken glass fibers, especially if it involves removal of cladding around the entire circumference of the fiber.
It is undesirable to produce excessive light losses.
If loss thraugh the fiber is minimized, it is possible to use an inexpensive light emitting diode as a source of light, and to use inexpensive photodetectors and amplifiers to detect the amount of light being transmitted through the sensor. For this reason, it is desirable to make sensors with as little loss as possible when at the maximally transmissive end of their range (low residual light loss), but with as large a loss as possible due to a change in curvature (high sensitivity). Preferably, the loss should be a linear function of curvature, with the centre of the linear range being at the centre of range of the mechanical quantity (such as curvature or displacement) being measured. These requirements often cannot be met with known sensors, because parameters such as residual light loss and sensitivity cannot be varied independently. For instance, sensitivity to curvature increases as the length of the treated zone is increased, but so does residual light loss.
Harvill et al. (U. S. Patent 5,097,252) have described intrinsic fiber optic sensors with the upper surface of the fiber treated to sense bending of fingers and other body parts. Although a monotonic output is claimed, the range of which includes a straight (zero curvature) sensor, the output is not linear and the range is not centred about zero cuzwature. Danisch (U.S. Patent applications Serial No.
07/738,560 filed on July 31, 1991, and entitled Fiber Optic Bending and Positioning Sensor and Serial No. 07/915,283 filed on July 20, 1992, and having the same title, each naming Lee Danisch as the inventor, and further described in "Bend-enhanced Fiber Optic Sensors," SPIE: The International Society of Optical Engineering, L.A. Danisch, Volume 1795, 204-214, September, 1992, Boston, MA, USA? "Smart Bone," Final Report for Canadian Space Agency Contract 9F006-1-0006/01-OSC, L.A. Danisch, 24 pp., June, 1992; and "Smart Wrist," Final Report for Canadian Space Agency Contract 9F006-2-0010/01-OSC, L.A. Danisch, March, 1993) has described fiber optic sensors with a surface of the fiber treated to emit light at a side with a minimal loss of throughput by means first described in U.S. Patent 4,880,971, also in the name of Lee A. Danisch.
The Danisch prior art includes descriptions of linear responses for a wide range of curvatures, a response that drops off as a cosine function for bends in planes not in the plane of maximum sensitivity, and a range centred about zero curvature. Another feature is a light absorbing coating which reduces or eliminates extraneous responses, including nonlinearity. Control over the positioning of the centre of the range would open the possibility of mainly using the portion of the range with the highest light throughput (lowest loss), rather than that with the lowest throughput (highest loss) as taught in Harvill. This would be especially useful if the centre could be adjusted without affecting the residual ' Tight loss or sensitivity, or adversely affecting the strength of the fiber. The prior art does not teach.how this can be done. The prior art describes sensors for which it is possible to vary the length, width and shape of a single treated strip, or the depth of multiple notches. Danisch, (U. S. Patent filings above) describes long sensors "...formed by alternating lengths of fibers with an emission strip with lengths of fully clad fibers." However, it is not shown how this technique can be used to gain control over the placement of the centre of the linear range.
A complicating factor in the manufacture of treated fiber optic sensors is that if their response to curvature is maximum in a given plane due to treatment not including the entire circumference of the fiber, then it can be difficult to maintain a proper orientation of the plane of maximum sensitivity after treatment but before embedment of the fiber.
The main problem is the ease with which the fiber can twist about its long axis due to torques applied at any point along the length of the fiber. This is a problem for any fiber whose complete circumference is not treated, including fibers that are treated at both the top and bottom, thus having a response characteristic that does not distinguish upward from downward bends, but that distinguishes (through a cosine law) between up/down and left/right bends.
Another complicating factor in the design of many intensity-based fiber optic sensor system is the need for a return path and a means of reflecting or turning the light at the end of the fiber run.
To eliminate the need for a turnaround and return path, a coupler is often used at the measurement end, such that light can be injected into a single fiber with a reflector at the end. Injected light travels through the treated portion of the fiber, is reflected at the end, and returns to the coupler in the measurement system in the same fiber. The coupler is designed to extract the return light only, passing it on to.a photodetector and amplifier.
Unfortunately, the coupler introduces large losses and can be expensive to manufacture. Also, the reflective structure at the end can be lossy and difficult to manufacture.
In other cases, i't is acceptable to use a return fiber with a reflective structure placed near the ends of the sensor and return fibers, whose distal ends face or are inserted into the reflective structure. Such a solution generally involves light leaving and re-entering the fibers, so that the sensor is no longer an intrinsic one, or it involves losses that may be unacceptable. It also invariably involves a reflective structure that is larger than the diameter of a single fiber or even two fibers, and is thus unacceptable for embedment.
A disadvantage of a turnaround loop at the distal end of a fiber optic sensing system is that even if sufficient width is available for the turnaround, it requires adding extra length to the system'beyond the location of the sensor.
This increases the size of the system and prevents sensing at the distal end of the system. For instance, it may be desirable to measure changes in curvature at the top end of a non-hinged but flexible lever which is being used as a "joystick" form of input device for a computer. If a turnaround loop is used for a fiber that enters the lever at the bottom, it would normally be at the top of the lever, thus not allowing known forms of curvature sensors to be placed at the top. As another example, if a turnaround loop is used and , it is desired to measure curvature at the centre of a curved beam, then the beam must be long enough to accommodate the turnaround.
Another disadvantage of the turnaround is that it must be held in position to avoid changes in light intensity due to changes in curvature within the plane of the turnaround, particularly if the turnaround has a small radius of curvature which is producing light losses substantially greater than those of a straight fiber. If the turnaround is 2~~~~.~3 rigidly affixed to the substrate, this may produce stresses on the fibers between the turnaround and the location of the sensitized zone, which must also be rigidly attached to the substrate in order to properly sense its curvature.
However, if the disadvantages of the turnaround can be overcome, it has overwhelming advantages in terms of cost of manufacture, small size, lack of complexity, and relatively low light loss.
The present invention provides an improved sensor means for sensing curvature and displacement with minimum manufacturing cost and minimum damage to the~fiber.
An object of the present invention is to provide a sensor means which minimizes residual light loss while optimizing sensitivity and preserving the strength of the 15w fiber.
A further object of the invention is to provide a sensor means which allows maximum utilization of the. portion of the linear range exhibiting the greatest transmission of light through the fiber.
A further object of the invention is to provide a sensor means that allows achieving a given residual light loss and sensitivity over a range of sensitized zone lengths.
A further object of the invention is to provide a sensor means which allows placing the sensitized zone of the sensor near the distal end of the sensor system.
A further object of the invention is to provide a sensor means which maintains the orientation of the sensitized zone, once treated.

~i3~~~

SUMMARY OF THE INVENTION
This invention achieves these and other results by providing a treatment means that minimizes residual loss and maximizes the effect of mechanical curvature of the sensor on the transmission of light through the fiber.
In one embodiment of the present invention, the sensor zone is made up of alternately treated and untreated portions of the fiber. There is opportunity for many light rays to refract toward the core from the untreated spaces between the treated ones so that average curvature may be sensed over a long length of the fiber without introducing unnecessary residual loss. The treated portions may involve modification of the fiber around its entire circumference or a part of the circumference. Treatment may include abrasion, 15' etching, repeated notching within the band, heating, chemical removal, and others. Treatment is generally accompanied by application of at least a thin layer of light absorbing material.
In another embodiment of the present invention, treatment of portions of the fiber mentioned above involves modification of only a portion of the circumference, these portions being oriented on the side of the fiber that is concave outward over a desired range of curvatures. By varying the lengths of treated and untreated strips, and the number of strips, and the extent of the circumference treated by each strip, the size of the linear region of sensitivity for concave bends can be increased, so that the sensor is .
operated near its maximum throughput condition. This ~~~8~z~~

embodiment has the advantage of allowing operation of the sensor within its highest linear throughput range, and controlling the size and placement of the centre of that range, while maintaining linearity of response to bending and displacement. The method has the advantage of being able to produce a sensor that has a minimum amount of circumference treated, while still providing maximum sensitivity and throughput over a desired range. This is important to maintaining the strength of the fiber used., especially when glass fibers are being treated.
In another embodiment of the present invention, the ' turnaround at the distal end of the fiber sensor loop is treated to be sensitive to bending, such that the treated portion has a minimal effect on throughput of the turnaround, 15' but the sensitivity to bending is the same or improved over that of a sensitized zone placed on a straight portion of the fiber. In this case, the net throughput of the sensitized zone and the loop combined can be made greater than for a sensor fiber in which the zone and loop are separated. The sensitized turnaround loop makes it possible t'o sense curvature at the free end of a structure, and has the added feature that if the turnaround loop is heat-formed or constrained by a form or fixture, it forms a plane that is always orthogonal to the plane of maximum sensitivity of the sensor, so that it is easy to maintain orientation of the maximum sensitivity plane during manufacture. The structure also lends itself to easier manufacture, because the turnaround forms the distal end of the fiber loop, and can ~~~,'a.r~x) easily be inserted into a machine for heat treatment, embossing, sanding, or other operations. This is not the case for a fiber which must be treated at some arbitrary location along its axis, especially if it must be inserted into an oven for heat treatment, without involving the leads in the treatment. This embodiment can be used with various forms of treatment, including treatment of various portions or all of the circuinfererice.
The invention may be defined as a fiber optic curvature and displacement sensor comprising a fiber optic light guide having at least one light emission surface extending, for part of the length of the guide, in a direction selected from:
a substantially circumferential direction, and a substantially curved axial direction when in a curved portion of a curved guide.
In one preferred embodiment, the light emission surface is in the form of a plurality of circumferentially-oriented bands.
. In another preferred embodiment, the light emission surface is in the form of at least one ring around the circumference.
Preferably the light emission surfaces are positioned to give the most desirable orientation to the maximum sensitivity plane.
The fiber optid light guide may be in the form of a loop, the loop having at least a substantial portion of the light emission surface or surfaces therein. These latter surfaces preferably are in a configuration selected from circumferentially-oriented bands grouped on an inside or concave portion of the curvature of the loop, and axially-oriented bands substantially parallel to the plane of the loop and substantially following the curvature of the loop.
In a preferred embodiment, from three to five circumferentially-oriented bands are grouped together with the bands each extending for about 60° to about 90° of circumference. In the case of. axially-oriented bands, preferably from two to six curved bands are either coplanar or in parallel planes. The invention includes sensors in which the fiber optic light guide, or portion thereof, is in the form of a single fiber which, serves both to illuminate the light emission surface or surfaces, and to collect illumination which has passed the light emission surface or surfaces, the illumination end and collection end of the fiber being located in a single region removed from the light emission surface or surfaces.
The invention includes, in a vehicle suspension .
system which includes an electronic system for sensing displacement between a vehicle body or frame and a vehicle wheel system and comprises a suspension sensor, the improvement comprising a flexible beam and a fiber optic curvature and displacement sensor mounted to the flexible beam, the sensor comprising a fiber optic light guide having at least one light emission surface extending, for part of the length of the guide, in a direction selected from:
a substantially circumferential direction, and a substantially curved axial direction when in a curved portion of a curved guide.
The invention also includes a method of sensing curvature and displacement, of an elongate member, comprising attaching a fiber optic light guide to the member, the light guide having a light emission surface extending, for part of the length of the guide, in a direction selected from:
a substantially circumferential direction, and a substantially curved axial direction when in a curved portion of a curved guide, the light guide extending along the member; injecting a light beam into one end of the light guide, detecting the light beam at the other end of the light guide, measuring the difference in the light beam between the one end and the other end, indicating curvature or displacement, of the member.
Preferably the method includes selecting the light guide so as to optimize sensing of curvatures over a range that includes, as a substantial portion of the total substantially linear range sensed, curvatures that produce increasing transmission of light with increasing curvature.
Preferably the method includes attaching a plurality of fiber optic light guides to the elongate member, each having a plurality of light emission surfaces such that the planes of maximum sensitivity of the guides are at different angles from each other so that at least one guide maximally indicates curvature at a unique planar inclination.

,!s'~
~r ~ r3 t~ ...

A preferred arrangement may utilize two to six fiber optic light guides oriented in different directions at predetermined angles relative to each other.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 is an overall view of a bending sensor apparatus, with the sensor shown cemented to a bending beam.
Figure 2A is a side view of a fiber with alternating bands of treated and untreated material.
Figure 2B is a cross section through one of the untreated portions of the fiber in Figure 2A.
Figure 2C is a cross section through a treated portion of the fiber in Figure 2A.
Figure 3A is a side view of another type of treatment, showing alternating bands of treated and untreated material, the treatment not encompassing the entire circumference of the fiber.
Figure 3B is a cross section through one of the untreated portions of the fiber in Figure 3A.
Figure 3C is a cross section through one of the treated portions of the fiber in Figure 3A.
Figure 3D is a cross section through a fiber as in Figure 3A, but treated in an alternate manner Figure 4A is a longitudinal cross section of a straight section of fiber with three emission surfaces as in Figure 3A, showing ranges of light rays subtended by emission surfaces;

Figure 4B is a longitudinal cross section of a downward curving fiber as in Figure 4A;
Figure 4C is a longitudinal cross section of a upwardly curving fiber as in Figure 4A;
Figure 5 shows a family of curves showing light loss through a fiber treated with successively more emission surfaces, as it is bent over a wide range of curvatures;
Figure 6 shows a more detailed graph of a fiber sensor treated as in Figure 5, to have a linear region centred on zero curvature as well as data from another type of sensor;
Figure 7 shows a sensor system including a loop of fiber used to return light to the optoelectronic measuring system. The loop is treated to act as a sensor and is mounted to the bending beam to measure curvature near the end of the beam;
Figure 8A is a longitudinal section of the loop in Figure 7. The top portion of the cladding in the loop has been removed to allow light to escape in two patches on a side of the fiber. The patches are located on curved portions of the loop;
Figure 8B is a plan view of the fiber in Figure 8A;
Figure 8C is a cross section through one of the treated portions of the fiber in Figure 8A;
Figure 8D is a cross section through one of the 2~9~~53 Figure 8F is a cross section through a treated portion of another embodiment like the fiber shown in Figure 8A except that the treated portion is located on the inner portion of the loop and is in a substantially circumferential orientation;
Figure 8G is a plan view as in Figure 8B, except that the emission surface is continuous;
Figure 9A shows_a plan view of a turnaround loop as in the embodiments portrayed in Figures 8A through 8G. Two rays which originate as rays in the plane of the loop are shown travelling through the fiber, remaining in the said plane;
Figure 9B shows a cross section through the fiber just before it begins to curve into the loop, showing the two rays still in the said plane;
Figure 9C shows a cross section through the apex of the loop, showing the position of the two rays, which are still in the said plane;
Figure l0A shows the same loop as in Figure 9A but with a ray that is parallel to the plane of the loop but displaced vertically;
Figure 10B shows a section through the fiber just before it enters the curve of the loop, showing the position of the ray above the plane of the loop;
Figure 10C shows a section through the fiber, containing the ray, showing deflection of the ray downward as it refracts from the outer wall of the loop;

20~~~5~
Figure 10D shows another section of the fiber, containing the ray, showing deflection of the ray again downward, at such an angle that is lost from the fiber at the outer wall of the loop;
5 Figure l0E shows sensor loss due to curvature for a fiber loop with two emission surfaces, and for a similarly treated straight section of fiber;
Figure 11A is a plan view of a loop treated to sense curvature of a diaphragm, attached along the surface of a 10 diaphragm;
Figure 11B is an elevation view of the sensor of Figure 11A;
Figure 12A is a plan view of a sensor as in 11A, except the loop is supported in a curve from above the 15 diaphragm and touches the diaphragm at a point, so as to sense displacement of the diaphragm;
Figure 12B is an elevation view of the sensor system of Figure 12A;
Figure 13A is a plan view of a loop treated to sense 20 curvature, attached at its apex to a rotating drum or shaft, for the purpose of indicating position of the shaft according to a varying curvature imposed on the loop;
Figure 13B is an elevation view of~the sensor structure of Figure 13A;
Figure 14 is a transparent view of a joystick input device containing loops of fiber treated and arranged to sense displacement of the flexible joystick handle in two degrees of freedom;

Figure 15A is a plan view of a loop treated to sense curvature, mounted as a cantilever beam for sensing vibration and acceleration;
Figure 15B is an elevation view of the sensor of Figure 15A1 Figure 16 is a side view of a spring and fiber loop system designed to translate large displacements of a body into relatively smaller displacements of the end of the loop;
and Figure 17 is a schematic diagram showing light paths and electronic circuitry.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Figure 1 illustrates a curvature or bending sensor 10 mounted with adhesive on bent beam 11. In the example, light is conveyed from a light source 12 through a plastic or other optical fiber light guide 13 to the sensor portion 10, and thence through guide 14 to a photo-detector 15. The light guide near the sensor region 10 has had its protective jacket 20~ (a layer or layers surrounding the cladding) removed and the core/cladding interface has been modified according to one of various methods of treatment described below. Portions 13 and 14 leading to the sensor region may have the jacket in place.
The sensing portion 10 is adapted to sense bending. the photometer 12 and photo-detector 15 are part of an optoelectronic measuring system 16 and display 17.
Figure 2A illustrates a fiber 1g with alternating circumferential bands of treated material 19 and untreated material 20. The figure shows the layers of the fiber out to the level of the cladding 23. The bands are formed by modifying the care and or cladding so that light is lost from the core. The bands can be formed by deliberately removing cladding as by abrasion, melting, etc. or by displacement as by pressure or rubbing on the fiber, for example by a heated tool, depending on the particular form of fiber: The treated bands are covered over with light absorbing material 21 which provides mechanical strength and environmental protection where cladding has been modified, but whose primary purpose is to absorb all light exiting the fiber from the core 22, which is shown in Figures 2B and 2C. Without this layer, the fiber can still act as a light guide because air surrounding it will have an index of refraction lower than that of the core. The fiber may actually transmit more light if the cladding is removed and no light absorbing coating is applied, thus preventing it from properly measuring curvature. In any case, without an absorbing coating the fiber will exhibit a nonlinear response that varies over time, especially if reflective materials, liquids, and dirt are present. With absorbing coating the response of the fiber is very constant over time and is unaffected by environmental factors. The absorbing coating may consist of carbon-filled epoxy, dyed elastomer, carbon-filled hot glue, or any other substance that permits light to exit the fiber but prevents it from re-entering in any substantial quantity. The light absorbing coating may serve other functions as well, such as protection of the fiber against environmental contamination. It may be 20~~1~~

applied only to the sensor portion of the fiber or may be incorporated into a part, such as a rubber or graphite/epoxy part, into which the sensor is embedded.
Figure 2B is a cross section of fiber 18 through X-X
where no treatment has been performed. Cladding layer 23 covers core material 22. Little light can escape from this section because the index of refraction of the cladding is less than that of the core. Excess absorbing material (not .
shown in this figure) may for convenience or for structural reasons cover the cladding in any of the untreated portions of the fiber.
Figure 2C illustrates a cross section of the fiber 18 through Y-Y, a treated portion. Where the cladding has been removed or modified, it is replaced by absorbent material 21, which covers the core 22 and any remaining cladding (not shown in this figure)'.
Figure 3A illustrates a fiber 24 with cladding 23, that has been treated as in Figure 2A except that the treated portions 25 cover a partial extent of the circumference.
Figure 3B shows a cross section through Z-Z, an untreated portion of fiber 24. This untreated portion is identical to that in Figure~2B and includes a core 22 and a cladding 23.
Figure 3C shows a cross section through P-P, a-treated portion of fiber 24. The cladding has been modified or removed over part of the circumference and back filled with light absorbing material 21.

Figure 3D is a cross section through a fiber treated like the fiber in Figure 3A but with both upper and lower surfaces treated to emit and absorb light. Although it is treated on both sides, it too will be maximally responsive to bends in the plane of maximum sensitivity passing through the centres of the two treated arcs. Its response to bends outside this plane will be a cosine function of the angle between the plane of maximum sensitivity and the other plane.
This cosine law is what allows the use of multiple sensor fibers with different angles of planes of maximum sensitivity to that three dimensional bands can be resolved by a set of sensors. It applies to any of these sensors except ones that includes bands or rings which extend the entire circumference. The best cosine effect happens if up to half 15~ the circumference is treated to form a band. Otherwise, there is a "DC offset" added to all the cosine responses up to the point that the signal does not change with different plane angles - this happens when the bands are completely circumferential in extent.
Figure 4A shows in longitudinal cross section a cone of light with an angle Ct that represents the solid angle of light that meets the internal refraction criteria for a fiber such as fiber 24 in Figure 3A., This angle is called the acceptance angle of the fiber. The size of Ct is determined by the relationship between the indices of refraction of the core and cladding.
The acceptance angle can be related to the indices of the core and cladding through ct - 2 sin's ( (niz - nzz) Biz) where the argument of the inverse sin function is called the numerical aperture of the fiber, which is determined from the index n~ and nz of the core and cladding respectively. Rays leaving representative point 26 with angles included in Ct will be refracted back into the core wherever they strike the cladding, and continue propagating down the fiber. Rays within angular ranges C1, C2, and C3 will strike the treated portions 21 and be lost.
Figure 4B shows angular ranges C4 and C5 within total cone angle Ct for the fiber 24 in a bent state. Because of the bend, the treated portions subtend smaller angles of light, so that C4 < C1, C5 < C2, and there is no C6 corresponding to C3 in Figure 3D. This is the mechanism which 15' produces an increase in light level as the fiber is bent to make the treated portions become more concave. Conversely, as it is bent in the other direction, the angles subtended by the treated portions become larger and represent a larger portion of the total acceptance angle Ct. This causes a loss of light compared to a straight fiber.
It is possible to increase the range of curvatures over which the fiber will have a substantial change in throughput for bends making the treated portions more concave.
This can be done by adjusting the axial lengths of the individual light emission surfaces and the length of the entire treated zone. Figure 5 shows the results of tests of a fiber treated with successively more emission surfaces along an axial line of the surface of the fiber. Each surface was 3 2~~~1~3 mm long, and spaced from the others axially by 3 mm. Data are shown for 2, 4, 6, and 8 emission surfaces. The fiber was held over various round mandrels to produce the curvatures shown. Curvatures include positive (emission surfaces concave) and negative (emission surfaces convex) values. From the family of curves in Figure 5 it is evident that when there are fewer surfaces, the positive curvatures tend to produce changes in light that are not linear with curvature, but the negative curvatures tend to produce linear changes. This effect is very evident in the 2.x 3 mm curve. As more surfaces are added, the linear portion of the curve moves .
toward the positive curvature region, as in the 8 x 3 mm curve. Intermediate values (as in the 6 x 3 mm curve) can be chosen to plane the linear range approximately intermediate between positive and negative curvatures.
Figure 6 shows a more detailed graph of a fiber sensor treated as in Figure 5 to have a linear region centred on zero curvature, as well as data from another type of sensor. The "centred" data are shown in the curve for "Treatment 1." The data for "Treatment 2" are from a similar fiber that was treated over its entire circumference so that it has a negative sloped response for positive curvatures.
Together, Figures 5 and 6 show that by applying various methods of treatment, it is possible to place the linear ranges of the fiber response curves so that they reflect various ranges of curvature, and to change the slope of curvature over a wide range.

The various curves shown in Figure 5 are explained in the following way:
For a fiber bent so as to make its emission surfaces more convex, (negative curvature in Figure 5), the surfaces continue to intercept more rays the farther the fiber is bent.
The fiber may be bent substantially in this direction with an ever increasing loss of light. However, the fiber will have a nonlinear response for very large negative curvatures (as seen in Figure 5, 8 x3 mm), because of the failure of the emission zones to intercept additional rays. As the fiber is bent the positive way, fewer rays are intercepted (see Figure 4C) as the bands take on an increasingly concave form. However, for a short emission surface, the surface is soon substantially out of the path of the rays. There is then no further loss of light, because it is predominantly reflected from the side of the fiber away from the emission surfaces, and does not interact with them. The interaction is least when there are few emission surfaces, as evidenced in Figure 5, 2 x 3 mm and 4 x 3 mm. The 2-surface fiber (and a 1x3 mm surface fiber which is.not shown) even shows an increased loss fox the largest curvature tested, probably due to microbending losses predominating over emission surface losses.
In one embodiment, the present invention utilizes such a large linear region for positive curvatures to increase the net throughput of the sensor system. By spacing the emission surfaces, it is possible to both increase the sensitivity and size of the linear region for positive curvatures and to minimize the residual light loss because a 2~~~~53 substantial number of rays can still pass the sensor region without attenuation. The extended structure permits distant hands to intercept light rays that are nearly parallel to the fiber axis, even when the curvature change is minimal. This leads to a high sensitivity with minimum residual light loss.
If one attempt to achieve high sensitivity for positive curvatures by lengthening a single emission surface, a limit is reached where the increase in residual light loss exceeds any gains in sensitivity, even though linearity is maintained.
For 1 mm plastic fibers, typical light emission surfaces for efficient sensors are approximately 2 to 10 millimetres in length, spaced by 2 to 10 millimetres. The overall length depends on the desired range but is typically up to 50 cm. Surfaces for smaller glass or plastic fibers are typically smaller. Notches may be used, but will not achieve the same sensitivity as uniform surfaces with smaller surface texture dimensions. This is probably because notches perform an emission function but also tend to scatter light back into the fiber. This is particularly true if they are not covered with a light absorbing layer. Notches have the further disadvantage of weakening the fiber.
If the emission surfaces occupy the total circumference of the fiber as in Figure 2A, there is no increase in throughput for bends of the fiber in any plane.
All of the sensing is done through decreases in light level.
Nevertheless, spaced emission surfaces are still an advantage for many sensors, as they can be used to sense average curvature over a greater axial length of the fiber. This can 2Q9~.~~3 eliminate or reduce undesirable effects from large local changes in curvature, for instance due to the presence of a foreign body under the fiber.
Ordinarily, a loop is used to return the light signal to the optoelectronic measuring system. Often, space is limited so that the loop must be formed in a tight curve such that substantial amount of light is lost from its outer convex surface. When combined with the residual light loss of a sensor elsewhere on the fiber, the resulting total light loss may be excessive. The "loop sensor embodiment" is designed to reduce this total loss by placing the emission surfaces in a novel manner.
In Figure 7 is shown a sensor system designed to have maximum throughput even though it includes a loop 28 at the end that may be iri a tight curve that loses light at its convex outer surface. This figure includes the same components as in Figure 1 except that the treated section 10 is on the loop 28 instead of on a straight section of the fiber. The sensor is designed to measure curvature of the substrate to which the loop is attached. In Figure 8, the sensor would be used to measure curvature at the end of the beam 11.
A detailed drawing of one embodiment of the sensor of Figure 7 is shown in Figures 8A through 8F. In Figure 8A, the loop is shown in longitudinal section. Emission surfaces 29 have been formed in one surface of the fiber by removing cladding 31 from core 30 and replacing it with light absorbing material 21. In Figure 8B, the emission surfaces 29 are shown in plan view. Cross sections 8C and 8D show sections through treated (D-D) and untreated (E-E) portions of the fiber respectively. The number of emission surfaces is representative only. Larger loops could contain more emission surfaces. Emission surfaces may be formed on surfaces above or below the plane of the loop, or on an inside or concave portion of the curvature of the loop. Figure 8E is a cross section illustrating emission surfaces on both the upper and lower surfaces of the loop. Figure 8F is a cross section illustrating an emission surface on the inner concave portion of the loop. Figure 8G shows another variation of loop sensor wherein the emission surface 29 covers virtually all of the upper surface of the loop.
Figure 9A shows the core 30 of the loop of Figures 8A through 8D. Rays h1 and h2 are in the plane of the loop.
They meet the outer convex surface of the loop and are refracted around the loop to continue on through the fiber.
For rays substantially in this plane, there will be little change in direction out of the plane as they traverse the loop.
Figure 9B shows a ray v1 that is substantially out of the plane of the loop but parallel to it. The ray is shown in Figure 108, a vertical (perpendicular to the page) cross section through A-A, just before the ray enters the curve of the loop. The ray impinges on the outer convex surface of the loop. Figure lOC shows a vertical section through the core containing the ray v1 after it refracts from this first collision. Because it is substantially above the plane of the ~U981~3 loop, it is deflected downward by the curve of the fiber and collides a second time with the wall of the core. Figure 10D
shows a vertical cross section through C-C, a plane containing the ray vl after its second collision. It is now travelling even more downward and impacts the outer curve of the loop at an angle such that it cannot be refracted back into the fiber, but travels through the cladding (not shown) and is lost.
Thus, rays that are travelling out of the plane of the loop but parallel to it will be deflected vertically as they travel through the loop. Rays above the plane will be deflected downward. Rays below the plane will be deflected upward. If they are sufficiently above or below the plane, they end up being lost because they strike the core/cladding boundary at too small an angle of incidence to the normal due to the quasi-spiral reflections indicated in Figures 10A
through loD.
We have seen above that there is a vertical impetus imparted to rays that travel through the loop without being lost at the outer convex surface of the loop. For convenience, these rays will be called "survivor" rays. They are distinct from "doomed" rays described Figures 11A through 11D that will either impinge directly on the emission surfaces early in their travel through the loop or be lost to the outer surface through excessive deflection. This change in elevation causes them to interact more or less with emission surfaces near the top or bottom of the loop. Thus, as the loop is curved out of its plane by an external stimulus, the emission surfaces interact with the survivor rays and cause the throughput to vary with curvature in much the same way as it does if the emission surface is located in a straight portion of the. fiber. By adjusting the length, spacing, and circumferential extent of the emission surfaces, it is possible to change the throughput so that it is linear with curvature and to adjust the midpoint of the linear range so that it includes the zero curvature point.
With the sensor on the loop, the emission surfaces can be adjusted so that a large proportion of 'the survivor rays interact with the emission surfaces. Those survivor rays that do not get absorbed by the treated emission surfaces for certain curvatures are by definition the rays that get through the loop. This leads to a high sensitivity of survivor rays to curvature. Survivor rays are made up predominantly of rays that enter the loop in planes near the horizontal (in the paper) plane of the loop and that have small vertical components. By contrast, "doomed" rays arrive in predominantly vertical planes with larger vertical components.
If instead of placing the emission surfaces on the curved loop, we placed them on the relatively straight fiber nearby, rays containing most of the sensor information would be in the vertical plane as they enter the loop and thus would be doomed rays. This would result in the light exiting the loop being made up mostly of rays that have not been modulated by the sensor. This is equivalent to reducing the sensitivity of the sensor. The same argument holds for a sensor downstream of the loop. The light leaving an untreated loop has been stripped of most of its vertical modes, so that a sensor placed in this light stream will be modulating a minority portion of the light passing through it. However, it is possible to form a useful loop sensor wherein the emission surfaces extend somewhat beyond the loop, as long as they are substantially on the loop.
Although the loop sensor embodiment has advantages of increased sensitivity and lower residual light loss for many arrangements of emission surfaces, this should not be taken to be its only advantage. Even in the absence of high sensitivity and low loss advantages, there are compelling reasons to form a sensor on the loop. These include sensing at the end of a structure, reduction of fiber length, ease of mounting, reduction of stresses on the fiber inherent in mounting a sensor portion and a loop portion separately with free fiber in between, ease of manufacture, preservation of orientation, and simplicity. In addition, loop sensors may be made to be relatively free from responses to bends within the plane of the loop.
The upper curve in Figure 10E shows the output of a loop sensor made by forming two axially oriented emission surfaces on the surface of a 7 mm diameter fiber loop of 1 mm diameter plastic fiber. The lower curve shows the sensor output when the same surfaces are formed on a straight section of fiber 2 cm from a 7 mm diameter untreated loop. For this arrangement of emission surfaces, the output of the loop sensor is superior to the output of a sensor formed apart from the loop.

209~~~3 Figure 11A shows another embodiment of the loop sensor, wherein a fiber 31, treated at the loop 32 to sense bending of the plane of the loop, is attached to a flexible diaphragm 33. These same parts are shown in Figure 11B. As the curvature of the diaphragm 33 changes, the output of the sensor changes. This sensor structure could be used to measure pressure or to form a membrane-type keyboard, or to perform many other tasks wherein a diaphragm undergoes changes in curvature.
Figures 12A and 12B show a sensor and diaphragm as in Figures 11A and 11B, wherein the fiber 31 is held in a fixed support block 37 so that at least the sensor portion of the fiber changes curvature according to the displacement of the diaphragm 33 at the point of attachment or contact 42.
Figure 13A shows a loop sensor wherein fiber 31 is treated at multiple portions 34 to be sensitive to bending of the plane of the loop. At the apex of the loop, it is attached to a turning shaft 35. Figure 13B shows an elevation of the same sensor structure, wherein it can be seen that the turning shaft can be turned over a limited angular range clockwise or counterclockwise about pivot 36 and that the loop is attached to an anchor point 37. As the loop winds around the shaft, its curvature increases, changing the throughput of light.
Figure 14 represents a joystick device wherein two loop sensors 38 and 39 are embedded in a vertical flexible shaft 40, attached to a base 41. The loops have planes of maximum sensitivity at right angles to each other, centred ". ~' .., . . ~'. W ' , ~.: . ,.. ... .'.' A..~ . :. ' ... ' .... ' , .
. '~ -' ': . .
' , ~ ' . ;. .. ' ' ~. ' , ' !. _ , ; , . ,; ' . _ . .

'.. ' ' :. .~, ~.~..' . . .' , : ~. ' ' ., ;.y ".: I~I; .. ... :': . .
.,..';.' .. :.~'. . .. ....~ , ;:, t :. . .. ~ ~_' ..
,. :.. ....: '~.; . ~ ,:;!~ ..,..~ .. - , .,. . .: .. ; - .. . . ' .. ~.

about the centre of the shaft. The outputs of the fiber sensors represent orthogonal components of the curvature of the shaft. In a miniaturi2ed form, this sensor could be a keyboard-mounted input device for a computer.
5, Figures 15A and 15B portray a loop sensor consisting of fiber 31 held by fixed support 37, with a sensor zone 32 on the loop. The structure to the right of the support 37 forms a cantilever beam. It is used to sense acceleration or vibration perpendicular to the plane of the loop. Optionally, a mass 43 may be attached near the end of the structure to modify the dynamic response of the sensor. The sensor could be used to perform a wide variety of acceleration measurements, including measurement of impact deceleration for the purpose of deploying an airbag protective system for automobiles. This structure is amenable to being manufactured by micromachining of a semiconductor or glassy substrate, wherein the fiber would be formed from the substrate and undercut by an etching process. A thin web or plate affixed to one side of the loop (not shown) could be added to prevent the shape of the loop from changing within its plane.
Figure 16 portrays a displacement sensor using a loop sensor 48 consisting of a.fiber 31 with a treated portion 32 on a loop at its distal extent, attached to a spring at-pivot point 44. The sensor is designed to sense a large displacement 47 of structure 46 relative to structure 45, which is fixed to a frame of reference to which the fiber is also fixed, bath by means of holding structures 37. By selecting the attachment point of pivot point 44, the movement 20~~:~~3 of the end of the sensor is restricted to a smaller range than movement range 47. The fractional amount of reduction is according to the patio of distance X2 to the total of distances X1 and X2, yet the reduced movement is linearly related to the movement of structure 46. The pivot point may take the form of a hinge, ball joint, flexible beam, cable, wire, elastomer, or various sliding contact points, or alternatively, the loop sensor may be mounted inside the spring or through the turns of the spring in various ways such that its curvature is linearly related to the linear movement of the pivot point. This embodiment allows the use of a small curvature sensor with limited travel to measure large displacements.
It will be seen that many of the sensors described are particularly convenient for being embedded in structures.
The loop sensors are particularly suited for measurements in thin tubes, pipes, rods and the like, particularly if measurements must be performed near the ends of the member.
The loop sensors are well adapted for use in various probes that must be inserted into small spaces. Applications for the sensors in general are meant specifically to include at least all applications that could potentially be performed with strain gauges, plus others.
The sensor may be incorporated into a means of transport, means for construction, agricultural implement, ~-robot, living body or a prosthetic device, for detecting a movement relative to a further movement or a point of 2~~~1~3 Figure 17 illustrates one example of electronic circuitry that can be used to measure the transmission of light through a paired sensor element such as that shown in earlier Figures. In Figure 17, fiber 19' (shown coiled to indicate arbitrary placement and length of the guide conveying light to and from the sensing portion) has the light emitting strip at 10°. Fiber 25', which is otherwise the same as fiber 19', has no sensing portion at position 26', which represents a section of the fiber in close proximity to sensor section 10. Both fibers are illuminated by photoemitters E1 and E2, which are light emitting diodes. Photodetectors D1 and D2, which receive light from the fibers, are PIN photodiodes, backbiased with 12 Volts to enhance the speed of their response to light energy. U1 and U2 are high input impedance operational amplifiers arranged as transimpedance amplifiers, converting light energy linearly into.voltages fed to the inputs of U3, which is an operational amplifier connected as a differential amplifier with a gain 10. The gain of amplifier U2 oan be varied with R1 so that for a straight fiber, the inputs to U3 are equal. In this condition, the optoelectronic circuit is analogous to a two-armed bridge such as is used to make strain-gauge measurements. Errors due to degradations in the fibers, connector variations, temperature fluctuations, and the like tend to cancel before reaching the output of U3.
The output of U3 is a voltage which varies with bending at the band portion 10. The output voltage can be further amplified and sent to a display unit or used to control various ~09~~~3 parameters, such as actuators designed to minimize the angle of bend.
Many variations of the circuitry are possible, including variations with greater immunity to error sources.
One such variation would use the same light source and detector, separating the signals by chopping them at different frequencies and employing synchronous detection. Another variation is to replace U3 in the above Figures with a divider circuit, so that the sensor signal is divided arithmetically by the reference signal.
While the above are descriptions of various preferred embodiments of the invention, various modifications should be obvious to those skilled in the art. For example, the sensor may be used with virtually any wavelengths or light including broadband or discrete spectra. Various referencing methods may be used, including a separate untreated fiber that is used to send light from the sensor light source over a similar path as the sensor fiber, the intensity of said light being used to perform differential or ratio compensation for common mode errors such as variations in fiber transmission or light source intensity. The reference path may also be used to automatically adjust the source light intensity to a fixed level. The invention is also meant to allow for emission surfaces that allow light loss at one wavelength or band of wavelengths but not at another wavelength or band of wavelengths. Thus, a two wavelength referencing method could be used over the same fiber, wherein light loss due to curvature is sensed with one wavelength and the other is used ~~~~i~~

to provide reference information for compensation,for common mode errors. Also, the light path may be chopped or modulated to provide improved sensitivity, if necessary. The output of the sensor may be used to measure parameters over a large range, or to determine a switching level. The sensors may be used for measurement, or as part of a control loop. The sensor fiber may be attached to a substrate that is undergoing bending, or the fiber alone, supported by two or more parts whose deflection is being measured, may be the element being bent. Consequently, the description should not be used to limit the scope of the invention.

Claims (29)

1. A fiber optic curvature and displacement sensor comprising a fiber optic light guide having at least one light emission surface extending for part of the length of the light guide and curved in a direction selected from:
(a) extending in a substantially circumferential direction, and (b) extending in an arcuate direction along the light guide when in a curved portion of said light guide.

2. A sensor as claimed in claim 1, having the light emission surface in the form of a plurality of circumferentially-oriented bands.

3. A sensor as claimed in claim 1, having the light emission surface in the form of at least one ring around the circumference.

4. A sensor as claimed in claim 1, wherein the light emission surface or surfaces are positioned to give a maximum change in light intensity transmitted through the light guide when said sensor is bent in a selected plane.

5. A sensor as claimed in claim 1, including means for injecting a light beam into one end of said light guide and means for detecting said light beam at the other end of said light guide.

6. A sensor as claimed in claim 5, including means for measuring the difference in intensity of said light beam between said one end and said other end of said light guide.

7. A sensor as claimed in claim 6, including display means for indicating any said difference in intensity of said light beam as a curvature or displacement of said light guide.

8. A sensor as claimed in claim 1, including a further fiber optic light guide positioned alongside said fiber optic light guide, said further light guide having an unbroken cladding layer and forming a reference light guide.

9. A sensor as claimed in claim 1, having said light emission surface or surfaces covered by a light absorbent material.

10. A sensor as claimed in claim 1, comprising a fiber optic light guide in the form of a loop, the loop having said curved light emission surface therein.

11. A sensor as claimed in claim 10, wherein the loop has a plurality of light emission surface regions therein.

12. A sensor as claimed in claim 11, wherein the surface regions in the curved loop are in a configuration selected from (a) circumferentially-oriented bands grouped on an inside or concave portion of the curvature of the loop; and (b) curved axially-oriented bands substantially parallel to the plane of the loop and substantially following the curvature of the loop.

13. A sensor as claimed in claim 12 (a) having three to five circumferentially-oriented bands grouped together, with the bands each extending for about 60° to about 90° of circumference.

14. A sensor as claimed in claim 12(b) having from two to six curved bands either coplanar or in parallel planes.

15. A sensor as claimed in claim 10, wherein the light emission surface or surfaces are positioned so as to give the most desirable orientation to the maximum sensitivity plane.

16. A sensor according to claim 1, wherein the sensor is a suspension sensor comprising a flexible beam and said fiber optic curvature and displacement sensor mounted on said flexible beam.

17. In a vehicle suspension system which includes an electronic system for sensing displacement between a vehicle body or frame and a vehicle wheel system and comprises a suspension sensor, the improvement comprising a flexible beam and a fiber optic curvature and displacement sensor mounted to said flexible beam, said sensor comprising a fiber optic light guide having at least one light emission surface extending for part of the length of the guide and curved in a direction selected from:
a substantially circumferential direction, and a substantially curved axial direction when in a curved portion of a curved guide.

18. A sensor according to claim 1, incorporated into a means of transport, means for construction, agricultural implement, robot, living body or a prosthetic device, for detecting a movement relative to a further movement or a point of reference.

19. A method of sensing curvature and displacement of an elongate member comprising attaching a fiber optic light guide to said member, said light guide having a curved light emission surface extending for part of the length of the guide in a direction selected from:
(a) extending in a substantially circumferential direction, and (b) extending in an axial direction along a curved portion of a said light guide, said light guide extending along said member; injecting a light beam into one end of said light guide, detecting said light beam at the other end of said light guide, measuring the difference in said light beam between said one end and said other end, indicating curvature or displacement, of said member.

20. The method of claim 19, including attaching a plurality of fiber optic light guides to said elongate member, each having a plurality of light emissions surfaces, said light emission surfaces on each individual light guide beam positioned to give a maximum change in light intensity transmitted through said individual light guide when said elongate member is bent in a unique selected plane.

21. The method as claimed in claim 20, utilizing two to six fiber optic light guides oriented in different directions at predetermined angles relative to each other.

22. The method as claimed in claim 19, including attaching a further fiber optic light guide alongside said fiber optic light guide, said further fiber optic light guide acting as a reference light guide.

23. The method of claim 19 which includes selecting the size, shape and orientation of each of the emission surfaces of the light guide so as to optimize sensing of curvatures, including curvature changes that produce increasing transmission of light with increasing curvature.

24. The sensor of claim 1 in which the fiber optic light guide, or portion thereof, is in the form of a single fiber which, serves both to illuminate the light emission surface or surfaces, and to collect illumination which has passed said light emission surface or surfaces, the illumination end and collection end of the fiber being located in a single region removed from the light emission surface or surfaces.

25. The sensor of claim 24, wherein the single fiber forms a narrow loop, at least a portion of the light emission surface is in the loop, and the apex of the loop is at the distal extremity removed from the illumination and collection region.

26. The sensor of claim 1 in which at least a portion of said fiber optic light guide is in the form of a single fiber which serves both to illuminate said at least one light emission surface and to collect illumination which has passed said at least one light emission surface, the illumination and collection end or ends of the fiber being located in a single region remote from the light emission surface or surfaces.

27. The sensor of claim 26, wherein the single fiber forms a loop, at least a portion of the light emission surface is in the loop, and the apex of the loop is at the distal extremity remote from said single region.

28. A fiber optic bending and position sensor comprising a fiber optic light guide in the form of a loop having a tight curve portion, with at least one curved light emission surface substantially on said tight curve portion, the size and positioning of the light emission surface being selected to increase sensitivity to deformation, means for injecting a light beam into one end of said light guide and means for detecting the light beam after it has passed the light emission surface zone.

29. The sensor of claim 28, wherein the light emission surface covers only part of the circumference, is axially-extending and covered by a light-absorbing material, the tight curve portion forms a semi-circular shape, and a deformation to be sensed tends to deform or deflect the semi-circular shaped portion.