SENSOR

Description

[0001] This invention relates to a sensor, and particularly, but not exclusively, to a sensor for sensing the position and / or shape of a soft robotic device.

[0002] Such a sensor has application in the field of minimally invasive surgery (MIS). MIS is being used with increasing frequency as an alternative to open surgery because of the improvements it can bring to patient safety and recovery time.

[0003] Surgical applications that benefit from MIS include reconstructive surgery, haemostasis, ischaemia detection, achalasia, gastric balloons / bands, catheters, balloons for heart bypass, auto-bionics. These, and other, applications may thus benefit from soft robots enhanced with sensors of this type.

[0004] Since the advent of laparoscopy, surgical technology has advanced on an exponential scale that has broadened the accessibility of the surgeon to the operative field with minimal incisions. MIS is usually carried out through natural body openings or small artificial incisions, usually resulting in reduced patient trauma, shortened hospitalisation, improved diagnostic accuracy and therapeutic outcome.

[0005] However, currently, the instruments used in MIS are rigid and, in the case of endoscopes or laparoscopic instruments, are difficult to control due to their long relative length and ergonomically unnatural controls. As the tools are also rigid, their use results in risk of unintentional damage to soft tissue which negates from the theoretical advantages of MIS techniques.

[0006] In the field of soft robotics, compliant materials are used to construct robotic devices that deform easily, making them highly applicable to interactions with soft tissue in MIS.

[0007] The next generation of surgical robotic devices will be designed to achieve access to all parts of the body from remote entry points, meaning they must be flexible and capable of navigating highly complex paths - all without causing damage to soft tissues.

[0008] A problem with soft robotic devices is that they are difficult to model, as both the structures of soft robots and the methods by which they achieve locomotion or actuation of joints rely on deformable materials. These materials exhibit non-linear responses to strain, so predicting the shape of soft devices and controlling the end effector is challenging, especially when external forces are exerted on the device.

[0009] Control would be facilitated if the bending angles of a soft device could be accurately measured. However, current shape sensing methods have limitations in terms of accuracy, robustness, compactness and deformability.

[0010] It is known to use electromagnetic (EM) tracking to track single points in space. However, a problem with this known technique is that the measurements are influenced by stray EM fields from other equipment. This makes shape reconstruction impossible and makes it difficult to integrate into an operating theatre with other electrical equipment.

[0011] In addition, it is known that imaging techniques such as fluoroscopy require large doses of radiation or contrast agents and either a C-arm or a monoplane X-ray system, while others such as ultrasound have low resolution and impair the flow of a surgical procedure.

[0012] It is also known to use optical tracking methods to track objects. However, such techniques require line of sight to the device and are therefore impractical or unsuitable in MIS applications.

[0013] Soft strain gauges made of Eutectic Gallium Indium (EGaln) have been used to sense contact, strain and shear deformation of the surface to which they are affixed. However, EGaln is expensive and lacks the biocompatibility important to MIS applications. Furthermore elastomer strain gauges give just a single degree of freedom and are either complex and costly to manufacture or too bulky to use in the restricted spaces imposed by MIS.

[0014] Fibre Bragg gratings (FBG) have been used to measure strain experienced by soft robotic devices. However, the location of a localized deflection cannot be determined, and the measurements are not sensitive in all directions. Therefore, a large number of sensor elements would be needed along the length of a device to fully sense the shape.

[0015] Electrical Impedance Tomography (EIT) is a sensing technique that infers variations in electrical impedance of electrically conductive media within a certain domain from voltage measurements made by electrodes positioned at the boundary of that domain.

[0016] Typical EIT techniques involve the placement of multiple electrodes positioned around the boundary of a domain such as a human chest or a chemical reaction chamber. An electric current is then injected between a pair of the electrodes and the resultant voltages that pass throughout the domain are measured by each of the remaining electrodes. This process is then repeated by sequentially injecting currents through different electrode pairs and eventually an image of the domain can be constructed based on the impedance inferred throughout the domain from the various measurements made by the distributed electrodes.

[0017] US 2014 / 0365009 A1 discloses a conductive skin formed of a flexible polymeric material that can be doped with a chemically inert material, such as a carbon-containing material. The skin can be wrapped around a housing. The skin can also include an array of electrodes for which can function as boundary-based tomographic units for sensing pressure distributions over the skin.

[0018] WO 2009 / 124209 A1 discloses tactile sensors are disclosed that mimic the human fingertip and its touch receptors. The mechanical components are similar to a fingertip, with a rigid core surrounded by a weakly conductive fluid contained within an elastomeric skin. The deformable properties of the finger pad can be used as part of a transduction process. Multiple electrodes can be mounted on the surface of the rigid core and connected to impedance measuring circuitry within the core. External forces deform the fluid path around the electrodes, resulting in a distributed pattern of impedance changes containing information about those forces and the objects that applied them.

[0019] US 2010 / 0049450 A1 discloses a measurement system that measures physical quantity distribution over an entire region using a sensor for physical quantity distribution (PQD sensor) in order to measure the physical quantity distribution that is regionally distributed in a two-dimensional or three-dimensional manner, as well as a method for measuring physical quantity distribution. The measurement system comprises: a body for converting physical multiple electrodes arranged in such a way that they can supply injection current to the PQDc body and measure data of the electric potential distribution generated by the PQDc body as a result.

[0020] According to the invention there is provided a sensor system comprising: a deformable structure; a plurality of electrodes operably connectable to the deformable structure; an electrical conductor operably connected to the deformable structure and electrically couplable to each of the electrodes; a plurality of current sources, wherein each current source is connectable to predetermined electrodes and produces an alternating current frequency different to those produced by the remaining current sources; a current injector for injecting current from each current source through predetermined electrodes; a voltage recorder for recording the voltage between predetermined pairs of electrodes; and an analyser for analysing voltage recordings to thereby determine the shape of the deformable structure.

[0021] The present disclosure provides a sensor system comprising a deformable structure. The deformable structure may be in the form of an actuator of the type known as a soft robot.

[0022] In the field of robotics, the term 'soft' refers to highly compliant materials that adapt to their surroundings in a similar way to those found in living organisms. Soft robotics is a subfield of robotics in which soft materials are used to construct robots, allowing for increased flexibility and adaptability in contrast to the rigidity commonly associated with standard robots. An important element of soft robotics is biomimicry, whereby the developers of soft robots are looking to the materials and actuation mechanisms present in nature for inspiration. One example is the way that plant cells utilise hydrostatic pressure to change the shape of leaves or petals in response to changes in the environment.

[0023] The system comprises an electrical conductor which is operatively connected both to the deformable structure and to a plurality of electrodes. The electrodes are each also operably connectable to the deformable structure.

[0024] A current may be injected through predetermined electrodes, and the voltage recorder is adapted to record the voltage generated between predetermined pairs of electrodes and caused by the injection of the current through the predetermined pair of electrodes. An analyser is then used to analyse the impedance recordings in order to determine the shape and position of the deformable structure.

[0025] The voltage recordings are analysed using electrical impedance tomography (EIT) techniques.

[0026] By means of the present invention, a soft robotic actuator may become proprioceptive (self-sensing) as well as being aware of its position relative to its surroundings.

[0027] In embodiments the current injector is adapted to inject current from the current source through a predetermined sequence of electrodes. In some embodiments of the invention the current injector is adapted to inject current from the current source through a predetermined sequence of pairs of electrodes.

[0028] In embodiments the current source is a source of constant current. In such embodiments an impedance dependent signal is created through the injection of the constant current.

[0029] In some embodiments the current source is connectable to some, but not all, of the electrodes. However, in other embodiments the current source is connectable to each of the electrodes. In such an embodiment, it is possible to inject a current through a plurality of predetermined pairs of electrodes at the same time.

[0030] Similarly, in embodiments the voltage recorder may be connectable to each of the electrodes. In such embodiments it is possible to record the impedance between a plurality of predetermined pairs of electrodes at the same time.

[0031] In embodiments the current source is connectable to a predetermined pair of electrodes, and the voltage recorder is connectable to each of the remaining electrodes.

[0032] In other embodiments of the voltage recorder is connectable to the predetermined pair of electrodes.

[0033] A current may be injected simultaneously to more than one pair of electrodes.

[0034] In such embodiments frequency division multiplexing may be used to enable more than one measurement to be taken and recorded at the same time. This means that more data may be obtained in real time. Such embodiments are particularly useful when used to determine the shape and position of a soft robotic device used in minimal invasive surgery, since it is possible to obtain real time, or near real time information about the position and shape of the soft robotic device.

[0035] Typical EIT systems inject current between sequential pairs of electrodes, known as Time Division Multiplexing (TDM).

[0036] However, the need within TDM EIT to switch between electrode pairs imposes a limit on the data acquisition rate.

[0037] By injecting multiple frequencies simultaneously through different electrode pairs Frequency Division Multiplexing (FDM) removes the need to switch entirely, making it a more suitable technique for real-time shape sensing. This is achieved by having a number of current sources working in parallel, each to inject an alternating current with a predetermined frequency through different electrode pairs. By recording multiple sets of impedance data simultaneously, an accurate reconstruction of the electrical conductor's shape can be formed instantaneously and then updated in real time or near real time.

[0038] The EIT method, and particularly the FDM EIT method, provides more information than the tracking of a small number of single points and does not require line of sight or a controlled environment, as in electromagnetic tracking and optical tracking. As such, this method is highly suitable to unstructured environments with limited space.

[0039] In embodiments the electrodes are electrically isolated from the surrounding environment.

[0040] In such embodiments, the electrodes can be electrically coupled to the electrical conductor in a manner that isolates the measurements from the surrounding environment, therefore, results can be obtained reliably. Furthermore, as the electrical signal is confined within the device, it is non-invasive and biocompatible.

[0041] In embodiments the deformable structure is an inflatable structure.

[0042] In such embodiments the electrical conductor may comprise an electrically conductive fluid forming part of the deformable structure. In particular, the electrically conductive fluid may fill the deformable structure. In such embodiments the deformable structure may be pressurised or otherwise inflated and / or depressurised or otherwise deflated by controlling flow of the electrically conductive fluid into and out of the deformable structural.

[0043] A particular advantage of such embodiments is that the deformable structure may be positioned in situ in a deflated condition. Once the deformable structure has been positioned appropriately, the electrically conductive fluid may be pumped or may otherwise enter the deformable structure in order to pressurise or otherwise inflate the structure.

[0044] In embodiments the sensor system further comprises a palpation sensor and the deformable structure forms part of the palpation sensor.

[0045] In such embodiments the plurality of electrodes and the electrica conductor may each also form part of the palpation sensor.

[0046] Palpation is a medical technique for diagnosis in which a practitioner would feel an area of a patient's body with his / her fingers or hands during a physical examination. The method therefore relies on touch rather than any other sense, such as sight or chemical sensing.

[0047] A palpation sensor (sometimes referred to as a tactile sensor or tactile palpation sensor) is a sensor that allows a practitioner to perform palpation techniques without using his / her fingers or hands, instead the sensor is able to 'feel' the area in question. Such a sensor may be particularly advantageous for use in MIS applications where a surgeon would typically feel an area of internal body tissue with his / her fingers or hands during an equivalent standard surgical procedure.

[0048] In such embodiments the deformable structure may be inflated with an electrically conductive fluid. In use, the deformable structure may be positioned against an area of internal tissue such that the deformable structure conforms to the surface it is positioned against. The change in shape of the deformable structure may be measured in correlation to a corresponding change of voltages recorded by the plurality of electrodes. Therefore the palpation sensor is able to 'feel' the shape of the internal tissue based on the deformation which is experienced by the deformable structure.

[0049] In embodiments the electrically conductive fluid comprises a saline solution.

[0050] In some embodiments the electrical conductor comprises an outer layer of the deformable structure.

[0051] In such embodiments the electrical conductor may be in the form of a skin enclosing some or all of the deformable structure.

[0052] An advantage of such embodiments is that the electrical conductor may be attached to an existing deformable structure.

[0053] In such embodiments the electrical conductor may comprise a fabric or material made from a conductive material.

[0054] In some embodiments the electrical conductor may also serve as an actuation mechanism for producing movement in the actuator.

[0055] For example, in embodiments in which the electrical conductor comprises an electrically conductive liquid, the liquid may also serve as a means for producing movement in the actuator. For example, by changing the pressure of the liquid within the actuator, movement such as bending may result. In such embodiments the conductive liquid acts as both the electrical conductor and a fluidic actuator. The movement that may be produced in such embodiments of the invention may include bending motion and variable stiffness.

[0056] In embodiments the sensor system further comprises a retraction device and the deformable structure forms part of the retraction device.

[0057] In such embodiments the plurality of electrodes and electrical conductor may each also form part of the retraction device.

[0058] In the field medicine, and particularly surgical medicine, a retraction device is a device for the retraction of a region of tissue to expose underlying tissue beneath it, thereby providing the surgeon with better visibility of the underlying area.

[0059] The electrical conductor of a retraction device according to an embodiment of the invention may serve as an actuator wherein actuation of the deformable structure causes a bending movement which may be used to retract internal body tissue.

[0060] In embodiments the sensor system further comprises a flexible printed circuit (FPC) wherein each of the plurality of electrodes is coupled to the FPC.

[0061] In such embodiments the FPC may comprise a flexible connection element which provides an electrical connection to the plurality electrodes.

[0062] A sensor according to such embodiments may readily be manufactured according to the following method: The FPC may be adhered to a first part of an outer skin such that a chamber is formed between the FPC and the outer skin, and the flexible connection element extends from the chamber. A laser welding process may then be used to seal the chamber by welding a second part of the outer skin to the first part in a given pattern such that the outer skin encases the FPC. The flexible connection element may be left unsealed at this stage before adhesive is used to make the outer skin watertight around the flexible connection element.

[0063] The flexible connection element provides a rapid connection to the plurality of electrodes that are now located in the interior of the inflatable chamber.

[0064] The invention will now be further described by way of example only with reference to the accompanying figures. Embodiments that do not fall within the scope of the claims are to be interpreted as examples useful for understanding the invention. Figure 1 is a schematic representation of a sensor Figure 2a is a schematic representation of a double hinged actuator forming part of the sensor Figures 2b and 2c are schematic representations of the double hinged actuator of Figure 2a expanding and actuating respectively, in response to pressurisation. Figures 3a to 3d are schematic representations illustrating how the actuator of Figures 3a to 3c is constructed; Figure 4a is a schematic representation showing the positioning of electrodes in the actuator of Figures 2a to 2c; Figure 4b is a schematic representation showing the current injection electrode pairs of the actuator of Figures 2a to 2c; Figure 4c is a graph representing how the impedance signal recorded between the current injection electrode pairs of Figure 4b varies according to the shape of the actuator; Figure 5 is a schematic representation of a palpation sensor forming part of a sensor Figure 6 is a graph showing changes in EIT voltages recorded over time in the palpation sensor shown in Figure 5; Figure 7 is a graphical model of the palpation sensor shown in Figure 5 when filled with an electrically conductive fluid; Figure 8 is a graphical model of the palpation sensor shown in Figure 5 when deformed by its environment; Figure 9 is a schematic reconstruction of the palpation sensor shown in Figure 5 which highlights the localised deformation shown in Figure 8; Figure 10a is a schematic representation of a double hinged pneumatic finger forming an actuator forming part of a sensor Figures 10b and 10c are schematic representations of the double hinged pneumatic finger of figure 10a shown expanding in response to pressurisation; Figure 11 is a schematic representation of a single hinged pneumatic finger forming part of an actuator forming part of a sensor Figures 12a and 12b are schematic representations of actuator forming part of a sensor Figure 13 a graph showing changes in EIT voltages recorded over time in the actuator shown in Figures 12a and 12b; Figures 14a and 14b are schematic representations of an actuator Figures 14c and 14d are schematic representations showing how the actuator of Figures 14a and 14b may be adapted to form a sensor Figures 15a and 15b are schematic illustrations showing the position of electrodes forming part of the sensor shown in Figure 14d; Figure 16a is a schematic representation of a target actuator shape; Figure 16b is a graphical representation of closed and open loop control schemes to achieve the target actuator shape shown in Figure 16a; Figure 17 is a schematic representation of a closed loop control system used to control the soft actuator of Figure 2a Figure 18 is a schematic representation of a closed loop control system used to control the soft actuator of Figure 14c.

[0065] Referring initially to Figure 1 a sensor according to an embodiment is designated generally by the reference numeral 2. The sensor 2 comprises an actuator in the form of deformable structure 4. Embedded into the surface of the actuator 4 is a plurality of electrodes 6. In this embodiment the actuator 4 comprises an inflatable structure comprising an outer skin 8 which is in the form of an electrical insulator. The actuator 4 comprises an electrical conductor which in this embodiment comprises a conducting fluid 10 which fills the actuator 4. The electrically conductive fluid 10 may be any suitable fluid, and in this example comprises a saline solution.

[0066] The electrodes are connected by conductive tracks and the electrodes and the tracks are formed in a conductive layer embedded in the outer skin of the actuator 4. The outer skin is shaped such that the electrodes and connector pads only are exposed.

[0067] This may be achieved by using any convenient process such as by using sputtering techniques, transfer or inkjet printer for example.

[0068] The sensor 2 comprises an EIT system 100. The EIT system comprises an alternating current source 112, a voltage recorder 114 and the plurality of electrodes 6. The EIT system 100 further comprises a means for addressing specific electrodes 6 in order to inject current between predetermined pairs of electrodes. In the illustrated embodiments, the means for addressing specific electrodes comprises a junction 116. The junction may be any suitable junction and may for example comprise a switch network which may be implemented by daisy-chained switch ICs or multiplexer ICs. This embodiment may be implemented using a manual plug board using jumper wires although this would be at the expense of configurability. The electrodes 10 are connected to the current source 112 and to the voltage recorder 114 by means of the junction 116.

[0069] In the illustrated embodiment the alternating current source 112 comprises four parallel current sources 118 each providing an alternating current having a frequency that is different to the frequency provided by all other current sources.

[0070] By means of the parallel current sources 118, a plurality (here 4) of frequencies may be injected simultaneously through different pairs of electrodes 10. This is achieved using frequency division multiplexing (FDM) which removes the need to switch. This makes the invention particularly suitable for shape sensing.

[0071] Although the illustrated embodiment comprises four parallel current sources, in other embodiments there may be more or fewer current sources, and in some embodiments of the invention there are six independent current sources 118. In such an embodiment each source can inject current at 165 µA between 10 Hz to 100 kHz between a single pair of electrodes, well below the IEC 60601 safety limits.

[0072] The voltage recorder 114 comprises a National Instruments (NI), USB-6216 16-bit DAQ at a 50 kHz sampling rate.

[0073] The current sources 118 and the voltage recorder 114 are connected to the electrodes 10 using a custom PCB.

[0074] To obtain a single EIT measurement, the voltage recorded on a single electrode is demodulated at a single frequency and averaged over several sign wave periods.

[0075] In embodiments the range of 3 ms to 250 ms was selected to take into account the limitations of USB data transfer speeds and DAQ buffer size.

[0076] The system also comprises an analyser 120 which may be in the form of a personal computer (PC) for example.

[0077] Each current source 118 is used to apply an alternating current across a predetermined pair of electrodes 10. The junction 116 allows the electrodes 10 to be individually addressed so that an alternating current may be applied across any two electrodes as necessary. The voltage recorder records the impedance across pairs of electrodes and the impedance information is analysed via an analyser such as a PC.

[0078] The sensor may be used with a wide range of deformable actuators.

[0079] Referring now to Figures 2a to 2c, a sensor according to another embodiment is illustrated. The illustrated sensor comprises a deformable hinged actuator 204 which, in this embodiment is a fluid actuated hinge adapted to bend at an angle predefined by the aspect ratio of a diamond weld pattern formed in a central portion of the actuator 204. The actuator 204 further comprises a plurality of electrodes 206.

[0080] In the specific embodiment described, the actuator 204 comprises a chain of two hinges 222 connected in series to produce a curling actuator having a chamber 209 segmented into three regions by the hinges 222, as shown in Figure 2c.

[0081] Referring now to Figures 3a to 3d details of a method of forming the actuator 204 is illustrated.

[0082] Initially, the number and positioning of electrodes 206, as well as the selection of EIT measurement protocol (the choice of current injection and measurement electrodes) requires optimisation for each application.

[0083] In the illustrated embodiment a Finite Element Model (FEM) is created replicating the geometry of the deformable structure with candidate electrode positions. A sensitivity analysis is made using EIT forward modelling software, which estimates the change in voltages during expected changes in shape. This information is then used to identify regions of low sensitivity or redundancy in the measurements, which is then used as feedback to update the electrode positions and measurement protocol for the next modelling iteration.

[0084] The forward problem in EIT is a combination of the quasi- static Maxwell equations with the addition of the Complete Electrode Model boundary conditions for a single pair of electrodes (1 = 1,2) ∇ ⋅ α ∇ u = 0 , σ δ u δ n = 0 off ∪ i = 1 x e i u + x i σ δ u δ n = V i , i = 1 , 2

[0085] Where σ is the conductivity, u is the electric potential, n is the normal to the boundary, e l is the lth electrode boundary, z l is the contact impedance on the lth electrode, and V l are the measured voltages. The finite element method is then employed to solve a discretized form of (1) for a specific domain. This is then repeated for all injection pairs and measurement electrodes. Typically, a linear relationship between voltage changes and conductivity changes is made, allowing for a construction of a Jacobian matrix J, which maps the changes in conductivity in each element to changes in voltage all electrodes: δ u = J δσ

[0086] However, as the aim is to reconstruct conductivity changes from voltages measurements (2) must be inverted. However, as J is ill-conditioned, regularization is necessary before it can be inverted.

[0087] Regularisation is then employed to limit the output to a realistic conductivity distribution. These reconstructions are performed on a coarser subdomain than the tetrahedral FEM in the forward problem.

[0088] Example regularisation schemes are Zeroth order Tikhonov, First Order Tikhonov, Total Variation and NOSER. The regularisation hyperparameter is chosen through objective means, such as cross-validation or L-curve: δσ ^ = arg min δσ δ v − f δσ 2 + λ δσ 2 J λ − 1 = J T J + λ I − 1 J T

[0089] Where δσ ^ an estimated conductivity change, δv are the boundary voltage change and λ is a hyperparameter controlling the extent of the regularization. The reconstructions were performed using the method developed by Aristovich et al., [K. Y. Aristovich, G. S. dos Santos, B. C. Packham, and D. S. Holder, "A method for reconstructing tomographic images of evoked neural activity with electrical impedance tomography spp. 1095-109, Jun. 2014.] with λ selected through cross-validation in a hexahedral subdomain.

[0090] Solid model approximations of the deformable structure are designed and used to create tetrahedral FEMs.

[0091] Once the location and positioning of the electrodes has been optimised, the layout of the electrode traces can be designed. This step involves two layers. The first layer is a highly conductive layer which is equivalent to the copper in traditional printed circuit boards and which forms the electrodes and tracks. The second layer is the insulation layer which ensures the electrodes and connector pads only are exposed. This could be achieved by a number of processes such as sputtering, transfer or inkjet printing.

[0092] In the case of the inflatable deformable structure 204 of the type shown in Figures 3a to 3c, the electrodes 206 are deposited onto sheets 224 which may be made from plastics materials. The electrodes are aligned with a weld pattern before being laser welded to form the outer skin 208 of the actuator 204.

[0093] As shown in Figures 3a to 3d, a chain of two hinges 222 is connected in series to produce an actuator 204. In this embodiment, the double hinge actuator 204 has a length of approximately 200 mm and a width of approximately 500 mm and a very low profile. The two diamond shaped hinges 222 measure approximately 20 mm by 10 mm. The hinges 222 are shown schematically in Figures 2c and 2d.

[0094] Figure 3c also shows the application of an IR absorbing dye which is deposited by for example a robotic arm using an adapted pen to mark out where the two sheets 224 should be welded together.

[0095] The finished product is then shown in Figure 3d after the two sheets 224 have been welded together. Any excess sheet material may be simply cut from around the welded device and fluid inlet tubes may be connected to the device. It is possible to incorporate features having sizes down to 1 mm due to the techniques used to create and weld the device together.

[0096] A plurality of hinges may be welded into a single chamber, a plurality of chambers may be welded in a single layer and also a plurality of layers may be stacked and welded together after an alignment step. A variety of designs for fluidic actuators with integrated EIT sensors according to embodiments of the invention may be produced using the described method, thereby enabling a very wide range of applications to benefit from embodiments of the invention.

[0097] The actuator 204 may then be placed in situ before being pressurised or otherwise inflated with a conductive liquid 210 which in this case is a saline solution.

[0098] Referring now to Figure 4a a side view of the double hinged actuator 204 is shown in more detail. The actuator 204 is shown in side profile has six electrodes 231, 232, 233, 234, 235 and 236 as shown in the figure and one chamber 209.

[0099] Given the symmetry of the design, there are three possible injection pairs.

[0100] For example, current may be injected at 2, 4 and 6 kHz between the pairs of electrodes shown in Figure 4b. Voltages may then be recorded between each pair of electrodes opposite to the injection pairs. In other words, recording between 234-235, 232-233 and 231-236 for injection between each of the pairs shown in Figure 4b. This means a total of nine measurements may be taken.

[0101] In this embodiment the regions of chamber 209 in the actuator 204 are independent.

[0102] The voltage measurements taken while current is being injected into the top two current injection pairs in Figure 4b are largely independent because of the concentration of current localised around the electrodes injecting current. When injecting between electrodes 233 and 236, negligible current flows into the first chamber region 209a, so this measurement is not sensitive to changes of shape of that chamber, or in the first hinge 222a. It is sensitive only to where the current flows, which is between the second 209b and third 209c chamber regions, and due to the design, it is most sensitive to the shape change in the second hinge 222b.

[0103] This means that a single multiple-channel impedance measurement that is sensitive only to the change of shape of the hinge of interest is possible for both the first hinge 222a and second hinge 222b respectively, as shown in Figure 4c. In Figure 4c, the channels 2, 4 and 5 show change in impedance, and therefore shape, of the conductor at first hinge 222a, while the channels 3, 7 and 9 in Figure 4c show change in impedance, and therefore shape, of the conductor at second hinge 222b. Channel 1 in Figure 4c shows a change in impedance of the conductor on the interior of the soft actuator due to changes in either the first hinge 222a or the second hinge 222b.

[0104] In this embodiment, the wires connecting the electrodes to the current source and voltage recorder may be in close proximity. This might cause capacitive coupling leading to interference between signals on different wires. The separation of the two injection frequencies may be increased to 10 kHz to minimise capacitive couplings. Capacitive coupling or crosstalk can mean that signals on one wire cause interference on nearby ones. By making the spacing between frequencies higher, it makes it easier to filter out any interferences which might occur. This was done in the case of the finger actuator 504 illustrated, because the wires used in the prototype had to be in close proximity.

[0105] Referring now to Figure 4c, changes during manual rotation of each axis in turn of the actuator shown in Figure 4a and 4b is illustrated.

[0106] It can be seen that there are large changes in the corresponding channels when the first hinge 222a is moved, but smaller changes when the second hinge 222b is rotated. However, for the injection across the whole sensor, changes are seen in every case.

[0107] Referring now to Figure 5, a palpation sensor 304 forms part of a sensor system according to an embodiment of the first aspect of the invention. The palpation sensor 304 comprises an outer skin 308 and a flexible printed circuit (FPC) 340 which, in turn, comprises a plurality of electrodes 306 and a flexible connection element 341 providing an electrical connection to the plurality electrodes 306.

[0108] In this embodiment the FPC 340 is adhered to a first part of the outer skin 308 such that a chamber 309 is formed between them. A laser welding process is then used to seal the chamber 309 by welding a second part of the outer skin 308 to the first part in a given pattern such that the FPC 340 is encased in the outer skin 308. The flexible connection element 341 is left unsealed at this stage before adhesive is used to make the outer skin 308 watertight around the flexible connection element 341. The flexible connection element 341, which extends from the chamber 309, provides a rapid connection to the plurality of electrodes 306 that are now located in the interior of the inflatable chamber 309. In this embodiment, there are sixteen electrodes 306 present, but there may be more or fewer electrodes present in other embodiments.

[0109] The chamber 309 may be inflated with an electrically conductive fluid (such as saline). In this embodiment of the invention, the FPC 340 further comprises an electroless nickel immersion gold (ENIG) surface finish which provides a sufficiently low impedance contact with the electrically conductive fluid (e.g. saline).

[0110] In use, the palpation sensor 304 may be used as a tactile palpation sensor which can be deployed endoscopically. Once inside the body, the chamber 309 may be inflated with electrically conductive fluid (e.g. saline) such that the palpation sensor 304 forms an inflated shape 411 shown in Figure 7. The palpation sensor 304 may then be pressed against a tissue of interest, which deforms the surface of the sensor such that the palpation sensor 304 forms a deformed shape 412, an example of which is shown in Figure 8 (although the deformation could cause any deformed shape). The change in shape results in changes in EIT voltages recorded by the plurality of electrodes 306. For example, the graph of EIT voltages shown in Figure 6 comprises stable periods 401 that indicate no deformation to the sensor and peaks 402 which indicate a deformation is taking place. The changes in voltages can then be reconstructed into an EIT image 420, showing the region of the sensor that was deformed, an example of which is shown in Figure 9. This is an improvement over existing pressure sensors used for palpation as it is not possible to localise the deformation on the pad surface of known palpation devices.

[0111] Turning now to Figures 10a to 10c, a deformable actuator 504 according to another embodiment is shown. In this embodiment the actuator comprises a finger actuator 504 comprising a pneumatic finger. The finger actuator 504 as shown in Figure 10a comprises a pneumatic beam 526 and two hydraulically actuated chambers 528.

[0112] The pneumatic beam 526 comprises two actuated hinges 522 formed from the two hydraulically actuated chambers 528. By increasing the pressure in, or otherwise inflating, the chambers 528, the pneumatic beam 526 is caused to buckle as shown particularly in Figure 10c.

[0113] The chambers 528 in the finger actuator 504 are each filled with an electrical conductor, in this case saline. Each chamber 528 is electrically isolated so current cannot be injected between the two chambers 528, nor can voltages be measured across electrodes 506 contained within separate chambers 528. Therefore, each chamber 528 can be treated independently. As each chamber 528 has only two electrodes 506 in this embodiment, there is only a single possible impedance measurement per chamber 528 - injecting current and measuring between that single pair of electrodes 506. The deformation of each actuated hinge 522 can therefore be analysed independently. A greater or lesser number of chambers 528 incorporating a greater or lesser number of electrodes 506 can be added to a pneumatic beam 526 or other deformable structure to create controllable hinges 522.

[0114] Referring now to Figure 11, a soft actuator 604 according to an embodiment of the invention is illustrated. The soft actuator comprises a pneumatic beam 626 and a hydraulically actuated chamber 628 coupled to the pneumatic beam 626. The chamber 628 comprises a plurality of electrodes 606 capable of EIT sensing.

[0115] Referring now to Figures 12a and 12b, a sensor system which may be used as a retraction device 603 comprises a pair of soft actuators 604 mounted to a flexible endoscope 629 such that the hydraulically actuated chambers 628 face away from the endoscope 629.

[0116] In Figure 12a the chambers 628 are empty and the soft actuators 604 therefore have a low profile which allows the flexible endoscope 629 to move unencumbered.

[0117] In Figure 12b the chambers 628 are inflated with a hydraulic actuation fluid, which also acts as an electrically conductive fluid, and the soft actuators 604 bend at the site at which they are coupled to the chambers 628. As the chambers 628 are positioned on the outward facing sides of the soft actuators 604, the soft actuators bend outwards, away from the flexible endoscope 629. The soft actuators 604 may be deployed in this way retract tissue away from a site on which a surgical procedure is to be performed, such as an endoscopy.

[0118] The plurality of electrodes may be electrically connected to an EIT system (such as that shown in Figure 1) to provide a measure of the shape of the soft actuators 604 and, therefore, provide a measure of the degree of bending of each actuator 604. In this way, a desired bending angle can be set, and feedback may be provided by the EIT system of the current bending angle. The pressure of the hydraulic actuation fluid can then be modulated to alter the bending angle until the desired value is reached. Failure to reach the desired value within a safe pressure range may also give an indication that external forces are acting on the retractor, thus preventing the desired motion.

[0119] Figure 13 shows a graph of EIT voltages recorded by a soft actuator 604 (similar to that shown in Figure 6). The closed periods 431, with EIT voltages of approximately 0 dV, indicate that the chambers are empty. Whereas the open periods 432 comprise large peaks of voltage the values of which correlate to a bending angle. Hence an EIT system can be used to inflate / deflate the hydraulic actuation chambers in order to achieve a set bending angle as set out above.

[0120] Turning now to Figures 14a and 14d, a soft actuator 704 according to another embodiment is illustrated. The soft actuator 704 comprises an existing soft actuator 750 which is adapted to deform in the manner shown particularly in Figures 14b, which shows the existing actuator 750 bending through an angle designated by the reference numeral 40.

[0121] The existing soft actuator 750 may be readily converted into an actuator suitable for use with a sensor according to embodiments by either introducing a conductive core 752 into the existing soft actuator 750 or by applying a conductive sheath 852 around the existing soft actuator 750 such that the conductive sheath surrounds the existing soft actuator 750.

[0122] Turning now to Figures 15a and 15b, an actuator 804 according to embodiments is illustrated schematically. The actuator 804 is shown to have an outer conductive sheath 852 with electrodes 806 embedded therein. The outer sheath 852 and electrodes 806 are shown more clearly in Figure 15b.

[0123] For an external actuator as shown in Figures 15a and 15b, the electrodes 806 will be placed inside the mould used to create the conductive polymer, thus embedding the electrodes inside.

[0124] Turning now to Figures 16a and 16b, closed loop and open loop control is shown.

[0125] The deformation of an actuator can be understood with the aid of kinematic models which calculate estimates of the actuator's shape based on EIT data.

[0126] Therefore, EIT data recorded by the voltage recorder 114 is analysed against kinematic models by the analyser 120 to produce an estimated reconstruction of the actuator's shape or deflection 40 in real time or near real time as represented in Figure 16a.

[0127] The reconstruction of the actuator is then used as a process variable in a control scheme.

[0128] A target shape 40 (Target X), and thus target tip position, is set and in one embodiment a closed loop control scheme is used whereby the automated controller iteratively controls the actuator and then evaluates the estimated shape of the actuator before providing a further control input. In MIS applications, variations in the environment are likely to cause continuous and varied deformation to the shape of the actuator. The EIT sensors measure any deformation and the shape reconstruction changes accordingly. The automated controller can then react to these changes near to instantaneously and continue controlling the actuator towards the target shape. The closed loop scheme therefore enables the target to be reached efficiently and accurately, as graphically represented in Figure 16b.

[0129] In a possible embodiment an open loop control scheme is used whereby the iterative feedback and control process is not performed autonomously but requires user input. The user is required to manually adjust the actuator in reaction to deformation caused by varying environments, this is represented by steps in the graph of Figure 16b wherein the flat sections result from lag between the estimated shape reconstruction being displayed and the user then reacting and providing further input accordingly. This control scheme is less accurate and less efficient, particularly in applications such as MIS.

[0130] Turning now to Figures 17 and 18, examples of the closed loop control schemes discussed in relation to Figure 16b are shown.

[0131] In one embodiment the closed loop control scheme shown in Figure 17 is used to control the soft actuator 204. A desired tip position is inputted into the system, it is compared with a tip position estimated by the EIT shape sensing system 100 by the error detector. The difference between the desired and estimated tip positions is sent as an error signal to the pressure controller which varies the pressure in the chamber of the soft actuator 204 to alter the actual tip position. The EIT shape sensing system 100 creates a new estimate of the tip position which is sent to the error detector and the process is automatically repeated until there is no longer any error between the desired tip position and the estimated tip position.

[0132] In another embodiment the closed loop control scheme of Figure 17 is used to control the soft actuator 504 in the same way as described for soft actuator 204. In a further embodiment the closed loop control scheme of Figure 17 is used to control the soft actuator 604 in the same way as described for soft actuator 204.

[0133] In one embodiment the closed loop control scheme shown in Figure 10 is used to control the soft actuator 704. Similarly, a desired tip position is inputted into the system, it is compared with a tip position estimated by the EIT shape sensing system 100 by the error detector. The difference between the desired and estimated tip positions is sent as an error signal to the movement controller which controls movement of the existing actuator 750 to alter the actual tip position. The EIT shape sensing system 100 creates a new estimate of the tip position which is sent to the error detector and the process is automatically repeated until there is no longer any error between the desired tip position and the estimated tip position.

[0134] In another embodiment the closed loop control scheme of Figure 10 is used to control the soft actuator 804 in the same way as described for soft actuator 704.

Claims

1. A sensor system comprising: a deformable structure (4); a plurality of electrodes (6) operably connectable to the deformable structure (4); an electrical conductor operably connected to the deformable structure (4) and electrically couplable to each of the electrodes (6); a plurality of current sources (112), wherein each current source (118) is connectable to predetermined electrodes and produces an alternating current frequency different to those produced by the remaining current sources (118); a current injector for injecting current from each current source (118) through predetermined electrodes; a voltage recorder (114) for recording the voltage between predetermined pairs of electrodes; and an analyser (120) for analysing voltage recordings to thereby determine the shape of the deformable structure (4).

2. A sensor system according to claim 1, wherein the current injector is adapted to inject current from each current source (118) through a predetermined sequence of electrodes.

3. A sensor system according to claim 1 or claim 2, wherein each current source (118) is a source of constant current, and the voltage recorder (114) is adapted to record an impedance dependent signal created through the injection of the constant current.

4. A sensor system according to any of the preceding claims, wherein the voltage recorder (114) is connectable to each of the electrodes (6).

5. A sensor system according to any one of the preceding claims in which the electrodes (6) are electrically isolated from the surrounding environment.

6. A sensor system according to any of the preceding claims, wherein the deformable structure (4) is an inflatable structure.

7. A sensor system according to claim 6, wherein the electrical conductor is an electrically conductive fluid (10) forming part of the deformable structure (4).

8. A sensor system according to claim 6 or 7, wherein the sensor system further comprises a palpation sensor (304) and the deformable structure (4) forms part of the palpation sensor (304).

9. A sensor system according to any one of the preceding claims wherein the electrical conductor comprises an outer layer of the deformable structure (4).

10. A sensor system according to any of the preceding claims, further comprising a controller for controlling locomotion or actuation of a soft robot according to the determination of the shape of the deformable structure (4).

11. A sensor system according to any of the preceding claims, wherein the sensor system further comprises a retraction device (603) and the deformable structure (4) forms part of the retraction device (603).

12. A sensor system according to any of the preceding claims, further comprising a flexible printed circuit (FPC) (340) wherein each of the plurality of electrodes (6) is coupled to the FPC (340).