Unit dose medication dispensing system

Abstract

A unit dose picking and dispensing system for use within a healthcare facility, comprising a plurality of storage modules 10 each configured to store a plurality of unit doses. A vision system is conf

Description

Filed of the invention The present invention relates to a unit dose medication dispensing system and method. More particularly, but not exclusively, the present invention relates to a picking and dispensing system and method for the location and retrieval of pharmaceuticals, medical devices, and medical supplies at the unit dose level from containers within in a healthcare facility, including a healthcare service provider. Background The effective management of pharmaceuticals, medical devices, and medical supplies in a healthcare facility, such a pharmacy, a hospital, and a healthcare service provider, requires multiple systems and processes to ensure the availability, safety, accuracy, and timely dispensing of the medication. Medication dispensing systems are used in hospitals, pharmacies, long term care centres, clinics, and the like to manage the medication storage and dispensing processes. Medications are first stored within the system and then located and picked on demand. A typical medication dispensing system includes a dispensing workstation and multiple storage bins or shelves. Each storage location is labelled to identify the medication, doses, or other medical supply held within. A healthcare professional would typically receive a medical prescription, read it, and then identify the items to be selected. The storage locations for each item will be identified, usually through memory or working through a common storage method such as alphabetical storage by name or grouping together of similar medications. The correct medication, in the correct dose, will then be manually picked and typically repackaged. Any relevant or important information regarding the dosage regime and the patient identify will be separately printed on a label and then manually applied to the repackaged medication. Finally, the medication is dispensed to the patient. This type of manual medication dispensing system is both cumbersome and prone to human error. Therefore, in recent years pharmacy automation has developed which is the use of technology and automated systems to control the medication dispensing process. These automated systems are designed to improve efficiency, reduce human errors, and reduce the burden placed upon the healthcare professionals. The most common type of automation is automated storage and retrieval systems (AS / RS) using robots configured to automatically store and retrieve medication on demand. However, automated dispensing systems for unit doses are highly specialised machines that are expensive to purchase, expensive to set up, and they are expensive to run. Automated dispensing systems also require trained professionals for initial set up and require daily maintenance. Furthermore, known automated storage and retrieval systems experience specific problems with the retrieval of medication. Many medications are supplied from the manufacture in original packaging. A common form of original packaging for individual tablet dosage forms is a cardboard box with blister packs comprising a plurality of medication receptacles formed in a transparent plastic sheet, arranged in an ordered manner, and reflective foil backing, for example aluminium foil, bonded to the backside of the transparent plasticsheet. Otherforms of medications comprise clear liquids that are packaged in reflective plastic and / or glass bottles, vials, or ampoules and metal caps contained in secondary packages (cardboard boxes). At the unit dose level, the inherent transparent and reflective nature of the primary packaging creates problems with the camera systems employed by the known automated systems because they cannot accurately locate and identify the medication. Further problems encountered by the known automated storage and retrieval systems relates to the large variety in the size and shape of medications, and medical devices, at the unit dose level. The automated system cannot handle such a wide variety in shape and sizes, thereby limiting their applications and effectiveness. Known automated storage and retrieval systems cannot recognise a unit dose once the medication has been segmented from the original packaging. Individual unit doses are also difficult to store in an ordered manner due to their size and shape. However, individual unit doses are more suited to being stored in a random manner because it makes the best use of the storage space and allows more products to be stored. Random storage is also quicker, and easier for the storer. Known automated storage and retrieval systems encounter problems with the recognition and identification of unit doses if they are stored in a random manner. Therefore, the known automated storage and retrieval systems require systematic repacking of the unit doses in standardised packages (such as pouches) prior to storing them and strict and ordered storage methods. This repackaging standardises the format of the unit dose and allows for simple picking by a robot arm. The location of the package is known when loaded in the system and retrieval is based on recalling that location. Optionally a barcode on the package allows for an additional verification. The repacking represents a significant additional cost to the healthcare facility requiring additional equipment and consumables. Furthermore, additional waste is generated in the medication dispensing process and there is additional work for the healthcare professional because every unit dose will need to be unpacked from before administration. With the increasing number of medications available in traceable unit dose format i.e., including a barcode on every unitdose in the original package, the repacking is a nuisance with a sole purpose to enable automated unit dose picking. There are less expensive automated storage and retrieval systems on the market in combination with pick-to-light and / or drop-to-light technology with medication stored at specific locations and display units, typically comprising LED lights, installed at those storage locations. The display units may flash or change colour to visually guide the user to the storage location. Drop-to-light systems allow the user to dispense the picked medication to the right patient / ward bins. Once the items are manually picked by the user, the lights can be manually changed by the user, or automatically change by activation of a sensor. However, these systems are still labour intensive and susceptible to human error. Therefore, a need exists to provide an improved automated medication picking and dispensing system and method for the efficient and effective dispensing of medication within a healthcare facility, especially for unit dose medication, with a high degree of automation / robotisation but without the need for systematic repacking. Summary of the invention It is an intention of the present invention to address at least some of the problems associated with the known automated medication picking and dispensing systems and methods. Drawings For a better understanding of the present disclosure, and to show more clearly how it may be carried into effect, reference is now made, byway of example, to the accompanying drawings, in which: Figure 1 is a front perspective view of an exemplary unit dose picking and dispensing system according to the present disclosure; Figure 2 is a top perspective view of an exemplary unit dose picking and dispensing system according to the present disclosure; Figure 3 is a side perspective view of an exemplary unit dose picking and dispensing system according to the present disclosure; Figures 4a, 4b, and 4c illustrate example acquired images according to principles of the present disclosure; Figures 5a, and 5b illustrate example acquired images according to principles of the present disclosure; Figure 6 illustrates an example fiducial marker according to principles of the present disclosure; and Figure 7 illustrates an example fiducial marker applied to a unit dose stored according to principles of the present disclosure. Detailed description Example embodiments will now be described in detail with reference to the accompanying drawings. The present invention provides an effective system that may be used for the location, selection, picking, and retrieval of a unit dose from a storage module. The system and method aim to provide a high throughput, improved accuracy, and reduced costs for the healthcare facility. The unit dose comprises a unit of a medication. The unit of medication may be a single dose or multiple doses. The medication is not particularly limited and may be any type of medication and includes medical devices and instruments. The unit dose can be in any form including individual pills, blister packs of pills, vials of a liquid, boxes of packaged medication, pouches, patches, cannisters or smaller medical devices such as inhalers and EpiPens. Once retrieved, the unit dose may be dispensed for distribution directly to a patient or for distribution throughout the healthcare facility. Figure 1 and Figure 2 show an example of a unit dose picking and dispensing system (1) according to the present disclosure. The unit dose picking and dispensing system (1) comprises a plurality of storage modules (10). The storage module (10) may be a discrete container, a removable module, a draw, a bin, or the like. The storage module (10) shown in Figure 1 comprises at least a base, and side walls such that they define an internal storage area in which a plurality of unit doses can be accommodated and retained. The internal storage area may itself be further divided into smaller storage areas. The plurality of stored unit of doses may the same type or they may be different types. The plurality of unit doses may be stored in an ordered manner, they may be stored in a random manner, or a combination of both. Preferably, the storage module (10) is configured to hold a plurality of unit doses in a random manner. Storage in a random manner is such that the unit doses are not stored in a structure or predefined way. In this way, the time taken to store the unit dose is reduced. Furthermore, the storage module (10) can effectively use the storage space available to maximise the number of unit doses that may be stored at the same time. Medication that is supplied in vials or ampoules may be stored in an ordered and / or formatted way such as with the use of labelled trays, which has the advantage of protecting the medication during storage, picking and manipulations. The storage modules (10) may be movable between an open configuration and a closed configuration and may comprise a lid and / or a locking mechanism to secure the contents and to prevent theft or abuse. The storage modules (10) can comprise identification means in order that they may be recognised and distinguished from each other. The storage modules (10) may comprise a label such that it can be visually recognised. The storage modules (10) may also comprise identification means such that it is possible to identify the medication that is loaded and stored within. The identification means may comprise details of the unit dose, or unit doses, such as the type of medication, strength of medication, use-by-date, manufacturer information, dosage form type etc. Storage module labelling systems can include at least one of fiducial markers, text labelling, RFID, or barcodes. Loading of the unit dose into the storage module (10) can be conducted manually or via a fully or semi-automated method, such as put-to-light. In this manner, the present invention may be integrated into prior established systems. The arrangement of storage modules is not limited and can include multiple discrete storage locations within a larger storage location. Figure 1 and Figure 2 show individual storage modules (10) forming part of a storage system within a larger storage location (11). The location of an individual storage module within the storage system can be allocated or tracked according to a medication dispensing system. The storage location (11) is preferably within a pharmacy or hospital. The individual storage modules (10) may also be loaded onto a racking system (12) as part of a storage system. Figure 1 shows an example of an arrangement of storage modules (10) on a rack or shelf forming part of a racking system (12) within a storage location (11) according to principals of the present disclosure. Individual storage modules (10) may be moved from a storage location and / or the racking system (12) to a rail system (13) and moved along that rail system in a predetermined manner. An individual storage module (10) may thereby be removed from the racking system (12) and / or a storage location and / or the larger storage location (11). An individual storage module (10) may then be returned to the racking system (12) and / or the storage location and / or the larger storage location (11) using the same rail system (13). This facilitates the removal and return of a storage module fortasks such as visual inspection, stock replenishing, stock removal or replacement, repair, and the like. In the arrangement shown in Figure 2 the system (1) further comprises a dispensing station (14) with a dispensing section (15) for dispensing a unit dose directly to a healthcare professional or patient. Preferably the dispensing station (14) comprises a plurality of dispensing sections (15). The dispensing section can hold a medication bin assigned to a patient. The unit dose picking and dispensing system (1) shown in Figure 2 also comprises a supply belt (16). The supply belt (16) provides for the supply and / or dispensing of at least one unit dose. The supply belt (16) can provide for receiving unit doses from the larger storage location (11) for storage and / or for dispensing or distributing dispensed medication throughout the healthcare facility. In this way the proposed system can integrate with existing standard intralogistics solutions such as conveyors and automatic storage and retrieval systems manipulating open bins holding the unit doses. In the arrangement shown in Figure 2, the unit dose picking and dispensing system (1) can use the racking system (12) to hold the storage modules (10) storing fast moving unit doses, while the supply belt (16) can be used to supply slow-moving unit doses. The slow-moving unit doses may be supplied from an associated automate storage and retrieval system (ASRS). The system can prioritise the picking and dispensing of the fast-moving medication while an associated ASRS retrieves the slow-moving medication. The system can then switch to pick and dispense the slow-moving medication as it is supplied into the system. After the system has picked a slow-moving unit dose the supply belt (16) will also allow for the unit dose to be placed back in the associated ASRS. A fast-moving medication is usually a common prescribed and / or repeat prescription medication, and a slow-moving medication is usually one that is infrequently and / or less frequently prescribed. In use, the picking and dispensing system can work on the fulfilment of multiple orders, or tasks, at the same time (multi-order picking and dispensing). For example, in a multi-order picking and dispensing mode, the system repeatedly picks fast moving unit doses from the storage modules (10) and dispenses them into the dispensing sections (15) of the dispensing station (14), or other sorting solution. Each dispensing section (15) can be associated with an individual patient and / or order. When a slow-moving medication (unit dose) is then supplied or presented to the system, the system can switch from picking and dispensing the fast-moving unit doses to picking and dispensing the slow-moving unit doses to complete multiple orders at the same time. In this manner the time taken for the system to complete multiple picking and retrieving tasks is reduced. The retrieval time for the most commonly requested fast-moving unit doses can be reduced. Thereby, the overall efficiency of the system is improved. In the arrangement shown in Figures 1, 2, and 3, the system further allows for the addition of medication, even medication not handled by the unit dose system, for instance by using drop-to-light LEDs, on a user side of the racking, rail, or dispensing sections, which is a side that can be accessed from outside the storage location (11). The time needed to stock check and to replenish a unit dose is also improved. In the arrangement shown in Figure 2 the system (1) comprises a robot (30). The robot may be a cobot (co-operative robot) designed for interaction with a shared space with a human. This allows human operators and healthcare professionals to operate in the same space as the robot (30). This also provides for the present invention to be easily installed and implemented within existing systems. The robot (30) is moveable between a plurality of the storage modules (10), the dispensing section (15), and the supply belt (16). The robot (30) automates the organization, storage and retrieval of medication to order. The robot (30) design is not particularly limited, and it can move in numerous ways, depending on the intended purpose and the environment. The robot may further comprise a transport unit which is a mechanism for transporting the robot. The transport unit can include wheels for transporting the robot in a transport direction, a motor for rotating the wheels, and the like. The robot may also comprise legs to move along a surface, including uneven surfaces. The robot may have aerial mobility, including propellers or rotors to provide lift and propulsion. The robot may also be connected to a rail system distributed across a designated area within the storage location (11), and thereby travel along an at least partially predefined path. In an example, the robot may be mounted on a rail and one direction of transport may be referred to as a positive direction, moving away from a starting position, and the other direction of transport may be referred to as a negative direction or return direction. The robot may also comprise a combination of types of mobility to complete retrieval of a unit dose efficiently and effectively. Referring to Figure 3, the robot (30) comprises a base (31) at its proximal end (32), the base (31) itself may be coupled to the floor or base surface of the storage location (11). Alternatively, the proximal end (32) of the robot (30), or the base (31), may be coupled to or comprise the transport unit. The robot (30) shown in Figure 3 comprises a main body which includes a robotic arm. A robotic arm is a mechanical limb able to move in along multiple axis and with multiple degrees of freedom. The degrees of freedom include, but are not limited to, rotational and translational movement along each axis. Preferably, the robotic arm is a 6-axis arm, with six axis of freedom, provided with infinite rotation and sufficient reach to pick an individual unit dose from an individual storage module. The robot is therefore able to adapt to the specific position and orientation of an individual storage and / or an individual unit dose. The robotic arm is configured to perform a wide range of movements within the storage location (11) and can move quickly and easily between the plurality of storage modules (10), the dispensing section (15), and the supply belt (16). The robot arm may have multiple joints (33a, 33b) designed to provide the multiple degrees of freedom to move in the variety of directions and orientations. In the arrangement shown in Figure 3 the robot (30) further comprises an end effector (34). An end effector is a device configured to interact with the environment. The end effector may be attached directly to the robot main body, or it can be attached to a distal end of a robotic arm. The end effector (34) can be any type such as impactive, ingressive, astrictive, or adhesive so long as it can retain and move a unit dose. The end effector (34) is preferably designed to grip a unit dose and to retain that unit dose during movement of the robot (30). The end effector may comprise a gripper, an ingressive mechanism, a mechanical tool, magnet, adhesive, or custom design for a specific task. The end effector may also be customised and optimized for a specific type of unit dose. The end effector may be capable of various grasping strategies to handle a wide variety of shapes and sizes. Multiple end effectors can be attached to the robot and / or robot arm to allow for the selection of the best suited end effector for each type of unit dose and / or unit dose pose in the container. Furthermore, the end effector is preferably configured to release a unit dose unharmed. The end effector may also be passive and rely on the action of the robot to affect an interaction with the unit dose. The end effector may be constructed from a single material, or it include a variety of materials, including metal, plastics, and fabrics. The end effector (34) is preferably a gripper, which allows the robot (30) to grasp and retain at least one unit dose. Preferably the gripper is a suction cup. This type of end effector is advantageous when used with a fragile unit dose. The suction cup may be oval or round and comprises flexible lips and bellows that allow the suction cup to conform to a variety in the surface shape of the unit dose. Preferably the suction cup contacts a flat side of the unit dose for improved stability when retaining and moving the unit dose. The suction cup may alternatively or additionally comprise a flat cup section comprising a sealing lip formed of an elastically deformable material. The flat shape and lower inner profile can allow for reduced pick up times and higher forces to be applied which is advantageous for larger unit doses. The end effector (34) may include at least one sensor to provide feedback and information on the surroundings. The sensor may be a tactile sensor to detect contact and pressure on the end effector. The sensor may be a camera, proximity sensor, temperature sensor, vacuum detector (air pressure sensor), or scanner. The sensor may be attached to the distal end of the end effector, which is the end closest to a target object. The system (1) may further comprise at least one scanner. The scanner may be configured to read a label, barcode, or RFID tag. The scanner can be located on the robot, on the storage module, or elsewhere within the storage location (11). Preferably the scanner is located on the robot and can scan a storage module, storage module contents, and / or unit dose. During the normal handling of medication by the picking and dispensing system, or more generally by the healthcare facility, blister packaging containing the medication will require sectioning by cutting to dispense an amount of a dosage form, or a unit dose, that is less than the amount originally supplied by the manufacturer. Therefore, the system (1) may further comprise at least one cutting device (not shown in the Figures). The cutting device can segment a unit dose of medication into a smaller unit dose of medication. Once a unit dose of medication has been segmented it may need to be re-labelled or have additional labelling applied. Therefore, the system (1) may further comprise at least one labelling device (not shown in the Figures). The labelling device comprises a printer, to print a label, and a transport unit, to transport the printed label and print medium. The unit dose can be labelled in order that it may be visually recognised and distinguished from surrounding objects. According to an aspect of the present disclosure, a unit dose comprises at least one marker. Preferably, the marker is a fiducial marker. Alternatively, or additionally, the unit dose may comprise at least one further marker such as a label, text, RFIS, or barcode. Where the supplied medication comprises bulk tablets and tablets in blister packs that are too difficult to cut into smaller unit doses they may be repackaged. Therefore, system (1) may further comprise at least one packaging device capable of applying packaging to a unit dose (not shown in the Figures). The packaging device may be in accordance with well-known repacking technologies and may be capable of applying a marker to the packaging. Some forms of medication may be supplied in vials or ampoules that are difficult to handle due to their fragile nature and shape. Therefore, the labelling device of the present disclosure is able to apply a label to a formatted tray, or receptacle in which the medication is stored, or directly to the vial or ampoule itself. The unit dose picking and dispensing system (1) comprises a vision system (not shown in the Figures). The vision system is configured to acquire an image of at least one storage module (10). The vision system my further be configured to scan and / or map the storage area of an individual storage module (10). The vision system may further be configured to scan and / or map the storage location (11). The vision system provides the ability to analyse, identify, and accurately locate objects within an environment, and especially within a storage module (10). The vision system may be located on the robot, and preferably on the end effector attached to the distal end of a robot arm. In this manner, when the robot is moved to a storage module, the robot arm and end effector can be positioned such that the vision system can capture an image an image of the entire storage area of the that storage module. Alternatively, multiple fixed vision systems can be used, one per pick location, allowing for the capture and processing of images while the robot and / or robot arm is not yet in place. This can result in faster cycle times for the system. The vision system is configured to acquire an image of the area of the storage module where unit doses are stored. Preferably the vision system acquires an entire image of a storage area of the storage module (10). In this manner, when a plurality of unit doses is stored within the storage area of a storage module (10) the vision system is configured to acquire an image of the plurality of stored unit doses. Preferably the system operates with an Al model to ensure optimal capture and instance segmentation of the acquired image. The acquisition of the image can be automatic, for example, once the system has ascertained that the vision system is a location such that it can acquire an image of the storage area, or it can be as a result of user input, for example, a user issuing a command. Once the vision system has acquired an image of a storage module it can issue a vision signal comprising the acquired image. The vision signal may only comprise the acquired image, or it may further comprise additional information regarding the storage module and / or the storage area. The issuance of the vision signal may be automatic, and preferably the vision signal is automatically generated once an image is acquired. The vision system includes at least one camera. The camera allows the vision system to capture images of the environment for analysis. The camera may be 2D or 3D camera. The camera may be located within the storage location (11) or on the robot (30). Preferably, the camera is located on the distal end of the robot (30), and further preferably the camera is located on the end effector. Once an image has been acquired by the vision system, the system (1) is then capable of analysing that acquired image to identify at least one unit dose. Image processing techniques can be applied to enhance the image. Analysis of the captured image is conducted by analysing means, preferably operating with an Al model. Machine vision algorithms can be employed to filter the results and allow for the selection of relevant information. In one example, object recognition of a unit dose comprises the use of machine vision powered by artificial intelligence. The Al model can be any Al vision model trained on storage module content images and unit dose images. The Al vision model can be trained on types of medications, dosage forms (for example, tablets, capsules, patches, liquids), medical devices (for example, EpiPen®, inhalers, derma roller) and packaging (for example, vials, ampoules, pouches, blister packs, boxes, and the like). The Al vision model can be a combination of Al powered segmentation such as the Segment Anything Model (SAM) released for general usage by Meta® in combination with traditional image processing techniques. Integration of the SAM model reduces the time needed to train the system on object recognition. In one example, the system performs 2D instance segmentation of the acquired image using SAM and maps the 2D pixels of the identified objects (segments) to a corresponding 3D pointcloud obtained through a 3D camera to locate the objects in 3D space. The 3D camera can be an RGB-D camera, or the like. Pretrained Al models may also be used to limit the amount of training needed, especially on first use or introduction of a new variant. The object recognition may be based on feature detection where features can be shape, size, edges, colour, known characteristics, surface decoration, and / or labels of unit doses or any feature identified by the Al system during training. Referring to Figures 4a, 4b, and 4c, the acquired image shows the storage area of an individual storage module (10). A plurality of unit doses is shown stored within the storage area of the storage module (10) and that plurality is thereby captured within the acquired image. As shown in Figures 4a and 4b, analysis of these acquired images allows for the identification of each individual unit dose from the plurality of unit doses stored in the random manner. As shown, an individual unit dose can be identified regardless of its position or orientation within the storage module, and even if it is partially obscured. Referring to Figure 4c, the storage area of a storage module (10) can be further divided into smaller internal storage areas (42) by internal walls (43). Each of the internal storage areas (42) contains a different type of unit dose, for example vials or small medical devices. Analysis can be conducted to identify each individual unit dose (40), of each type, within in each of the internal storage areas (42). Referring to Figure 5a and 5b, further examples of a plurality of unit doses stored in a random manner. In this example a unit dose comprises reflective and transparent packaging and clear liquid medication. As shown in Figure 5a, the acquired image does not easily distinguish between the individual doses. However, as shown in Figure 5b analysis of the acquired image by the system (1) allows for the identification of a plurality of individual unit doses despite the transparent and reflective nature of the unit dose and their random arrangement within the storage module. This is especially advantageous for medications which are often small and comprise transparent or semi-transparent liquids and transparent and reflective packaging. If analysis of the image does not result in the identification of at least one unit dose, and / or no useable information is retrieved from a captured image, then the system can control the movement of the robot to manipulate the contents of the storage module in order to change the location of the unit doses. This manipulation can comprise shaking the container, or blowing air while moving a robot arm around in the container, or any other movement that changes the position and orientation of the unit doses in the container. A further image of the storage module may then be acquired, and analysis of that image can be conducted. Once a unit dose has been identified, the system selects at least one identified unit dose to be picked and dispensed. If a single unit dose is identified, then that unit dose is selected. If a plurality of unit doses is identified, then the system selects one of the unit doses from that plurality. Preferably the system operates with an Al model to select a unit dose. Referring again to Figure 5b, an Al model identifies each of the identified unit doses as a possible picking candidate (50). Picking candidate (50) can be identified based on characteristics of a segmented image of the acquired image (features). Characteristics can include shape, dimension, colour or the like. If different types of unit dose are identified, then selection of picking candidates can further be based on the type of unit dose. A plurality of picking candidates (50) is identified in Figure 5b and therefore the analysis includes identifying the best picking candidate from that plurality. Analysis includes the consideration of the likelihood of a successful picking for each identified picking candidate. Identification of the best picking candidate may comprise analysis of the spatial information for each picking candidate, and can include the height position of the picking candidate within the storage module, the distance of the picking candidate to the internal surfaces of the storage module, and orientation within the storage module. A picking candidate at the highest position within the storage module may be the best picking candidate because it will reduce the time taken to pick, it may also be the easiest to pick requiring the least amount of movement by the robot. A picking candidate in the middle of the storage module may be the best picking candidate because it reduces the possibility of the robot hitting a side wall of the storage module. A picking candidate with a level orientation (for example, the most alignment with the horizontal direction) within the storage module may be the best picking candidate because it provides an accessible surface to which a suction cup end effector can interact. The best picking candidate may also be identified based on learning from prior picking and dispensing by the system (reinforcement learning). Once an identified unit dose has been selected for picking and dispensing, integrated software can plan an optimal path (or picking trajectory) from a start position of the robot to the selected unit dose. Preferably, planning the optimal path is conducted using an Al model. The optimal path avoids collisions between the robot and ancillary devices and other objects within the environment, and thereby improves efficiency and success rate. The system can further work to determine an optimal picking point on a unit dose. An optimal picking point will depend on the type of robot and / or end effector employed by the system. An optimal picking point will be one that is easily accessible to the end effector. For example, with a suction cup end effector, a good picking point is a sufficiently large flat surface (or convex surface with large enough radius) and with a normal of that surface being at an angle of less then 45° with the vertical direction. For a mechanical gripper end effector, a good picking point is also one that can withstand a constantly applied gripping force without risking damaging the unit dose. The co-originates of an identified optimal picking point are then used to plot the path of the robot (30), and preferably the end effector (34), to pick the selected unit dose. Analysis to identify an optimal picking point may also be used when selecting an identified unit dose for picking and dispensing. The system further comprises a controller configured to control movement of the robot to pick and dispense at least one unit does from at least one storage module. The controller controls movement of the robot along the optimal path to the picking point. The controller can form part of the robot (30) or can otherwise be in communication with the robot to effect control. The controller may be further configured to acquire the acquired image from the vision system. The controller may comprise means to analyse the acquired image or is in communication with means for conducting such analysis. The controller can include one or a plurality of ICs, non-volatile memory, and the like. The IC includes CPUs and / or GPUs as a processor, a ROM, a RAM, and the like. In the controller the processor, such as the CPU, executes arithmetic processing according to a program stored in the ROM, the other memory, or the like, while using the RAM or the like as a working area, to control movement of the robot. For example, the controller executes processing according to firmware, which is a type of the program. Note that the processor is not limited to the single CPU / GPU, and may be configured to perform the processing using a plurality of the CPUs or GPU's or a hardware circuit, such as an application specific integrated circuit (ASIC), or may be configured to perform the processing using the CPU in conjunction with the hardware circuit. Examples of the controller include a personal computer, a smartphone, a tablet terminal, a mobile phone, a server, or a device having a similar level of processing capability as those devices mentioned above. The execution of movement by the robot is realized by the controller. The execution of movement by the end effector may also be realized by the controller. The system can further allow feedback and corrections. If a picking attempt fails, the system can learn from the failed planned path, and picking point, and consider factors such as the type of unit dose, or other environmental factors. Thereby the system can improve accuracy, speed, and efficiency overtime. Furthermore, the position of the unit does within the storage area of the storage module may change as the unit doses are successfully picked and / or as the result of a failed picking attempt. If a picking attempt fails, the system can work to repeat the acquisition, analysis, and identification steps to select a new picking candidate (50) (reinforcement learning). Machine learning and Al may be further be integrated into the system to learn as the system operates and to adapt and maximize running procedures and efficiency. Object recognition, peaks running times, medication requirements based on wards, seasons and temperature, and recognition of new medications with special requirements can all be used by the system. The system is therefore able to improve overtime and continue to optimize results, optimize timings, adapt to new environments and medications, and reduce errors. The system is further able to prioritise a plurality of orders by calculating an optimal path for each order and executing the orders in an efficient manner. The movement of the robot and end effector, along with the securely retained unit dose, may then be controlled to dispense the unit dose. The unit dose can be dispensed to any desired location accessible to the robot. The robot may be controlled to place the unit dose in a predetermined location, such as a tray, the supply belt (16), a dispensing station (15), or human hand. The robot may also be controlled to dispense the unit dose to location convenient for the application of further packaging. According to a further aspect of the present disclosure, the unit dose can comprise a marker, and preferably that maker is a fiducial marker. The vision system is further configured to detect such marker. The means to analyse the acquired image can also recognise the marker and process information from that marker. At least one fiducial marker may be located on each unit dose, or a storage tray containing a unit dose, stored within a storage module (10). The fiducial maker library should include a large set of markers with unique coding. When a sufficiently large library of markers is used a different code may be applied to each unit dose thereby providing the ability for the system to check if the fiducial marker read by the vision system is correct for the expected unit dose. Errors in product identification can therefore be detected during the picking process. Figure 6 shows a label (60) comprising a fiducial marker (61) according to one example of the present disclosure. The type of fiducial marker is not limited but may include Aruco FM, Apriltag, and / or RuneTag. The label (60) may also contain further identification information such as a QR ordatamatrix code (62), medication name, medication dose, expiry date, batch information and serial number (63). The label (60) may be already applied to the medication supplied by the original manufacturer. Alternatively, the label may be applied before or during the process of storing the unit doses within the storage location (11). As already discussed, during processing of medication by the healthcare facility, the originally supplied medication may be segmented into smaller unit doses and the label (60) may then be applied to individual tablets cut from a blister packet. The labels are such that they do not interfere with existing datamatrix barcode technology and can be applied side-by-side with existing labels. Therefore, the fiducial marker will not cause confusion for a barcode reader and the neighbouring datamatrix label can be read. Figure 7 shows plurality of unit doses within a storage area of a storage module (10). Each individual unit dose includes a label (60) comprising a fiducial marker (61). The plurality of unit doses is shown arranged in a random manner. The fiducial marker (61) can be detected within the acquired image, and the fiducial marker (61) provides information to facilitate calculation of its precise position in the X, Y, and Z axis. The fiducial marker (61) further provides information that allows calculation of its precise rotation position (Alpha, Beta, Gamma). Therefore, when a fiducial marker is applied to a unit dose, that fiducial marker, and the unit dose with which it is associated, can be identified by the system, and the precise position and rotation in space, within the storage module, can be calculated. Additionally, when a marker is partially obscured it will not be recognised by the system and therefor it will not be identified as a picking candidate. This provides an advantage in that the system will not first try to pick unit items that are underneath other items that obscure the marker. A storage module fiducial marker may also be applied to the inside bottom surface of a storage module such that when the acquired image is analysed the storage module fiducial marker can be detected. When a storage module fiducial marker is detected an alert can be produced by the system to indicate that the storage module is empty or nearly empty. A side wall fiducial marker may also be applied to a further surface of the storage module (10), or an internal wall (43) of the storage module, such that detection of a side wall fiducial marker can aid identification and position of that storage module. The further surface of the storage module (10) may be a side wall or lid. The fiducial marker provides an advantage in that it can be detected within an acquired image even if the unit dose to which it is associated is small and / or comprises reflective or transparent packaging, or if the unit dose comprises a medication that is transparent or partially transparent, such a clear liquid. The fiducial marker can thereby aid the identification and localisation of a unit dose. The fiducial marker can also be the picking point, or a picking point can be at a known and fixed distance and orientation from a fiducial marker. The fiducial marker can thereby provide a reference to a picking point. The size of the fiducial marker (61) is preferably larger than the end effector (34) such that, when it is the picking point, it ensures a free trajectory for the end effector (34) to the fiducial marker (61). Furthermore, when the end effector is a suction cup, the fiducial marker can provide a flat surface to which the suction cup can attach. The system may further comprise at least one sensor. The sensor can continuously, or periodically, monitor the robot (30) and end effector (34) location and movement, and can provide feedback information on real time location during picking and dispensing. The sensor may be a camera, a light sensor, a sound sensor, a proximity sensors, a position sensors, a velocity sensors, a tactile sensors, a temperature sensor, and / or ora depth sensor to create a 3D representation of a storage module and contents. Preferably the system comprises multiple sensors which may be the same type or different types. During the reaching, picking, and retaining movements, sensors on the robot and end effector can provide feedback on the environment to optimize safety and efficiency of the movement. The end effector can be manipulated based on the feedback to increase the success rate of the picking and retention of the object. The system may further comprise a user interface to allow a user to input an order to retrieve a unit dose, monitor the system, provide in-task commands, and / or define order parameters. The user interface can comprise a touch screen and / or keyboard. The system may be further provided with a display unit for displaying visual information. The display unit can comprise, for example, a liquid crystal display, an organic EL display, or the like. The display unit can include a display and a drive circuit for driving the display. An operation accepting unit may also be provided and is a unit for accepting an operation by a user, and is realized by, for example, a physical button, a touch panel, a keyboard, a mouse, or the like. The touch panel can be realized as a function of the display unit. The display unit and the operation accepting unit can be referred to as an operating panel of the system. The display unit or the operation accepting unit may be a part of the robot or may be an external peripheral device coupled to the robot. The present invention involves the combination of hardware and software to perform at least a picking and dispensing process accurately and efficiently. An example of a unit dose picking and dispensing method according to the present disclose will now be described. The system (1) first acquires an order to pick and dispense at least one specified unit dose. In one example, the order corresponds to a single retrieval task to be performed by the system and comprises a request to retrieve a unit dose of medication held in a storage module (10) within the storage location (11). The order can be acquired by manual input by a user on the operating panel of the system. Alternatively, the order may be acquired automatically by scanning a medical prescription and processing character recognition. Alternatively, the order may be automatically transmitted to the system from an external device or system. Once an order has been acquired, the system identifies a storage module (10) within the storage location (11) storing the unit dose to be picked and dispensed. The system may know the location of the correct storage module based on a medication storage system, or the system may scan each storage bin in sequence until a storage bin storing the unit dose is located. Alternatively, the system may send an instruction to a coupled ASRS to retrieve a storage module with the requested medication and present it to the robot, e.g. via the supply belt (16). Once the correct storage module has been located the vision system acquires an image of that storage module. Preferably, the acquired image is an image of the storage area within the storage module. The acquired image is then analysed to identify at least one unit dose within the storage area. Image processing techniques can be applied to enhance the image. If a plurality of unit doses is identified, the system then determines which of the plurality of identified unit doses are picking candidates. Picking candidates comprise the specified unit dose and represent good options for picking. The system can also rank the identified options in order based on likelihood of successful picking. If required, an additional step can be added to the method whereby the robot is controlled to manipulate the contents of the storage module if no useable information is retrieved from a captured image. The step of acquiring the image my then be repeated. The system then selects a unit dose to be picked and dispensed. The selected unit dose is usually the one with the highest likelihood of a successful pick from the identified picking candidates. A robot (30) is then controlled to pick and dispense the selected unit dose. In one example method, controlling comprises executing a planned path of the robot from a starting position to the identified unit dose; controlling the movement of an end effector of the robot to reach, grasp and / or safely retain the identified unit dose at a picking point; and controlling the end effector to safely release the unit dose at a dispensing location. If required, an additional step can be added to the method whereby the robot is controlled to present a further side of the unit dose, comprising a fiducial marker or label, such as a 1 or 2 D barcode, to the vision system to verify that the unit dose corresponds to what was requested, ensuring higher levels of medication safety. This also allows the system to identify medication that may have been stored in the wrong storage module. In the above manner, the present invention can quickly and accurately pick and dispense medication to order. The automated process allows healthcare professionals to have less involvement in the dispensing of the medication, thereby reducing staffing costs, reducing human error, and allowing healthcare professional to prioritise tasks. The system is further able to improve throughput, serving more patients than existing systems, and manual or semiautomated systems. The configuration and method described above may be realized by a single system but may also be realized by a controller and a robot system that are communicably coupled to each other within a healthcare facility. Multiple systems can work in parallel or together for increased throughput and to compensate for redundancy in case of failure of a robot or component. The system can be deployed in central pharmacies or at ward level. It will be appreciated by those skilled in the art that although the disclosure has been described by way of example with reference to one or more arrangements, it is not limited to the disclosed arrangements and that alternative arrangements could be constructed without departing from the scope of the disclosure as defined by the appended claims.

Claims

1. A unit dose picking and dispensing system for use within a healthcare facility, the system comprisinga plurality of storage modules, each storage module configured to store a plurality of unit doses;a vision system configured to acquire an image of at least one storage module;a robot moveable between the plurality of storage modules and configured to pick and dispense at least one unit dose from at least one storage module;anda controller configured to control movement of the robot to pick and dispense the at least one unit dose based on analysis of the acquired image;whereinthe storage module is configured to store the plurality of unit doses in a random manner; andthe system further comprises means to analyse the acquired image to identify at least one unit dose stored by the storage module, and then select at least one identified unit dose for picking and dispensing.

2. A unit dose picking and dispensing system according to claim 1, wherein the vision system comprises at least one camera; and preferably wherein the camera is a 3D camera.

3. A unit dose picking and dispensing system according to claim 1 or claim 2, wherein the vision system comprises 3D perception technology.

4. A unit dose picking and dispensing system according to any preceding claim, wherein the robot is a collaborative robot.

5. A unit dose picking and dispensing system according to any preceding claim, wherein the robot comprises a robotic arm comprising at least one joint.

6. A unit dose picking and dispensing system according to claim 5, wherein the robot comprising an end effector; and preferably the end effector is located on a distal end of the robotic arm.

7. A unit dose picking and dispensing system according to claim 7, wherein the end effector comprises at least one grasping tool, and the grasping tool preferably comprises a mechanical gripper or suction cup.

8. A unit dose picking and dispensing system according to any preceding claim, wherein the robot further comprises at least one sensor configured to generate a sensor signal based on a robot position or an environmental factor, and preferably wherein the sensor comprises temperature sensor, light sensor, or an acoustic sensor.

9. A unit dose picking and dispensing system according to any preceding claim, wherein identification of at least one unit dose stored by the storage model is based on analysis from an Al model.

10. A unit dose picking and dispensing system according to any preceding claim, wherein the selection of at least one identified unit dose for picking and dispensing is based on analysis from an Al model.

11. A unit dose picking and dispensing system according to any preceding claim, wherein the controller comprises the means for acquiring the image from the vision system and the means to analyse the acquired image.

12. A unit dose picking and dispensing system according to any preceding claim, wherein the system is further configured to plan a picking path for the selected unit dose, and the picking path is planned using analysis from an Al model.

13. A unit dose picking and dispensing system according to any preceding claim, wherein the system is further configured to select a picking point for the selected unit dose, and the picking point is selected using analysis from an Al model.

14. A unit dose picking and dispensing system according to any preceding claim, wherein the unit dose comprises a fiducial marker.

15. A unit dose picking and dispensing system according to claim 14, wherein the fiducial marker provides a reference to a picking point for the selected unit dose.

16. A unit dose picking and dispensing system according to any preceding claim, wherein the system further comprises a labelling device for labelling a unit dose, preferably wherein the labelling device applies at least one fiducial marker to the unit dose.

17. A unit dose picking and dispensing system according to any preceding claim, wherein the system further comprises a cutting device for segmenting a unit dose.

18. A unit dose picking and dispensing system according to any preceding claim, wherein the system further comprises a packaging device capable of applying packaging material to the unit dose; preferably wherein the packaging material comprises a fiducial marker.

19. A unit dose picking and dispensing system according to any preceding claim, wherein the system comprises an Al model configured to learn from a successful robot picking and dispensing.

20. A unit dose picking and dispensing method comprising:receiving an order to retrieve a specified unit dose;identifying a storage module, from a plurality of storage modules, storing the specified unit dose;acquiring an image of the identified storage module; andcontrolling the movement of a robot to perform picking, from the identified storage module, and dispensing of the specified unit dose based on analysis of the acquired image;whereinthe specified unit dose is stored in the storage module in a random manner; and analysis of the acquired image comprises identifying at least one specified unit dose, and then selecting at least one identified unit dose for picking and dispensing.

21. A unit dose picking and dispensing method according to claim 20, wherein the method further comprises the step of dispensing the unit dose in a designated location; and preferably issuing a confirmation signal.

22. A unit dose picking and dispensing method according to claim 20 or claim 21, the method further comprises a step of learning from successful picking and dispensing, and then optimizing controlling the movement of a robot.

23. A unit dose picking and dispensing method according to any one of claims 20 to 22, wherein the identified storage module comprises a plurality of unit doses stored in a random manner.

24. A unit dose picking and dispensing system substantially as described herein with reference to, and as shown in, the accompanying drawings.

Images