Abstract
Mobile Robots, or more properly defined as Autonomous Guided Vehicles (1), are vehicles which have been equipped with a drive system under computer control which allows autonomous guidance between two locations. Mobile robots are finding diverse applications as rovers on distant planets, providing access to high radiation environments, and performing daily routines for the elderly. Eventually, even the daily commute to work will not require your mental participation.
The recent availability of sophisticating computing capabilities at the level of a computer board has created a revival of the mobile robot field after their introduction in the 1960s followed by a languishing market in the late 1980s. Initially, robots were used in industrial plants where they became an obvious choice for moving material over greater distances than conveyor belts could reasonably perform. In the 1960's the “Hopkins' Beast” prowled around the corridors of Johns Hopkins University looking for nourishment in the form of power from wall outlets. Other mobile robots have been trained to run down the hallway looking for open doors, climb stairs, swing like monkeys from Jungle Jim bars, or even swim around a swimming pool using a fish like tail (2).
However, relatively little progress has been made towards using mobile robots to deliver laboratory supplies or participate in the process of laboratory analysis. However, there is significant economic incentive to introduce mobile robots into laboratories. For example, the physical movement of medical specimens, pharmaceuticals, blood products, patient charts, x-rays, and meals costs over 1 million dollars annually for hospitals over 500 beds. In small hospitals, 84% of specimen transportation is performed by the most skillful members of the laboratory, trained medical technologists (3). Furthermore, a medical technologist has been estimated to spend 20% of his/her day moving specimens around the laboratory.
In buildings where laboratories are spread out over a large geographic area, specimen delivery can account for an even larger percentage of technologist time. Pharmaceutical laboratories process large numbers of samples which have a greater likelihood of being standardized to be compatible with mobile robots.
Many transportation options exist as commercially available products, for example, conveyor belts, pneumatic tubes, gimbled carts, and mobile robot. Conveyor belts are used widely in industry where production schedules are predictable and the temporal relationship between the input of raw material and output of finished products is predictable. However, many laboratories are not yet organized in a streamlined fashion because technologists must react to the changing numbers of laboratory requests, additional processing steps, and problem specimens. Providing conveyance to laboratories requires flexibility in order to accommodate the changing analytical requests.
Serial transportation, in which each specimen is treated as a discrete entity, allows specimens to be redirected to their analytical destination. Specimens can also be sorted into analytical batches which may be transported to analytical stations. As long as the batches are frequent and keep pace with the analytical instrument, the efficiencies of a batch process is similar to serial conveyance and in some cases will be more efficient. Programmable mobile robots (Table 1) are beginning to be used to provide a flexible transportation scheme for a few laboratories. These flexible devices are particularly useful in laboratories where there is a long geographical separation between the source and the analytical instrument. Conveyor belts may be too costly to install when distances begin to exceed hundreds of yards or are located in another building. A typical conveyor belt system from an industrial manufacturer costs over $350 a linear foot.
Mobile Robots
When designing conveyance systems the following points should be addressed:
Determine accessible pickup and drop-off locations Map the shortest path between pickup and drop-off points Determine the method for delivery (conveyors or mobile robots). Program the routing and scheduling of vehicles on the paths and at the pickup and drop-off points.
Hierarchical taxonomies for designing mobile robotic systems have been developed (4). For example:
A complete laboratory transportation system will undoubtedly become a combination of mobile robots, fixed conveyance and humans. Of these, mobile robots have the ultimate flexibility vs. cost for many facilities due to the ease in which routes can be changed. There is little doubt that the sophistication of mobile robots will accelerate over the next decade so that eventually this technology will be able to follow voice commands with flawless precision.
Mobile robots have resulted from the reduced costs of component parts and increased computing capacity of today's microprocessors. The basic elements of a mobile robot are drive motors, sensors, computer, batteries, and external case. Some mobile robots are not only capable of moving autonomously along the floor, but have also been programmed to ride elevators, open doors, and find their destinations while not running over patients. Robots aimed at the medical market require that they can be safely operated around the staff and patients. They must be able to perform routine tasks in the face of unpredictable obstacles such as pediatric patients and the elderly who might have limited visual and audioacuity.
Delivery must be performed as efficiently as possible with full notification if the assigned task cannot be carried out. An idle robot which has been disabled in a corridor containing vital patient specimens and information (x-rays or charts) could be disastrous for a critically ill patient. A disabled robot should notify the administrator of its difficulty. The power source of a mobile robot should allow charging during idle moments or require a battery change at predictable intervals or some combination of the two. Finally, the robots should be repairable from the factory via model or Internet or by a relatively technologically unskilled user on-site.
HelpMate (HelpMate Inc., Danbury, CN)
Size, Guidance Method, Sensor Technology
The HelpMate robot is 55″ high, and 36″ long and 32″ wide and weighs approximately 600 lbs. It is guided by following an electronic blueprint which has been down loaded into its computer memory (5). Reflective tape strips are also mounted on the ceiling at selected intervals so that the HelpMate can calibrate its wheel revolutions with the actual location in the building. The orientation of the robot within the hallway (Figure 1) is determined by measuring by sonar the distance to both walls, as well as comparing its internal gyroscope with the information obtained by other sensors. The HelpMate is covered with 24 sonar sensors which provide information about the environment around the HelpMate.
The HelpMate enters a hospital ward carefully avoiding the pediatric bassinet and infusion pump which has been placed in the corridor since the last HelpMate visit.
Facing forward and to the side are sonar sensors which, like the sonar used by bats, indicate the presence of an object with reflective properties. Infrared light transmitters are used on the top of the robot to communicate with door openers. In addition, the robot has a vision system which faces 45 degrees downward in front of the robot and covers an area between 18″ and 60″. If all else fails, the robot has collision bumpers mounted front and rear to prevent the 600 pound device from doing any damage. Emergency stop buttons are also located on either side.
When the HelpMate needs to change floors, it signals its request using a radio transmitter to a modified elevator (Figure 2). The HelpMate boards the elevator which has been determined to be empty. As soon as the elevator doors open, the HelpMate warns potential riders to “stand clear” before boarding. Hallway doors are opened by retrofitted mechanical devices which are commercially available and often installed for wheelchair access. The automatic door openers are triggered to open and close by infrared remote control signals initiated by the robot.
After the elevator reaches the designated floor, the HelpMate emerges to continue its deliveries. The elevator is returned to normal service after being vacated by the robot.
Alternately, the robot can signal an annunciator which will sound an alarm inside a closed door to announce its presence. HelpMate uses pre-recorded messages to deal with unexpectedly closed fire doors. However, the HelpMate makes use of its sophisticated obstacle avoidance software to do a remarkable job of avoiding getting stuck while en route.
The HelpMate is equipped with a transportation pack (Figure 3) with a rear opening door capable of carrying medical specimens, late food trays, pharmacy orders, blood transfusion products, medical records, radiology images, to name a few. The internal dimensions of the payload compartment is 18″ × 28″ × 26″ and can accommodate loads up to 100 lbs. The compartment can be simply latched or equipped with a combination lock to prevent theft of contents during the run.
Laboratory specimens are placed into the backpack of the HelpMate for delivery to distant locations.
RoboCart (CCRI, Inc., Lake Arrowhead, CA)
Almost human in shape, the RoboCart (Figure 4) is the smallest mobile robot on the medical market. The RoboCart is 41.5 inches high (3 feet, 6.5 inches high) and carries a specimen tray which is 18 inches wide and 26 inches long. An optional cover for the specimen tray brings the total height to 47 inches. The tray extends out over the width of the cylindrical base and allows for a one inch clearance over the top of a workbench. The ability to extend over the top of a counter allows for easy loading and unloading.
The RoboCart from CCRI (Lake Arrowhead, CA) is a light duty robot that can fit into the tight confines of many medical laboratories. Medical specimens are placed in the tray at the top of the robot which is also available with a locking cover.
Guidance is provided by 16 light emitting and light sensitive diode pairs which follow a reflective tape (3M tape) applied to the floor. Collision avoidance is provided by a single sonar sensor located on the front of the robot. The sensitivity of the sonar sensor is under software control and therefore can be adjusted to accommodate different conditions on the robot delivery route. Even the most crowded laboratories are accessible to the 2010 cart since it has a small footprint and can rotate on a central axis.
Instructions which guide the robot microprocessor come from a laptop computer which is contained inside a removable cover. The software is written in simple text files which consist of forward, left, and right commands. Motions are referenced to the guide path on the floor. Initiation of text files is made through a hand held wireless remote control operating on less than 5 milliwatts of power.
The RoboCart communicates with a host computer via a wireless local area network (LAN). The wireless LAN technology and scheduling algorithms will allow many 2010 Carts to be guided along a maze of intersecting paths.
Communications and Central Monitoring System
RoboCarts have a two level communication system. The lower level system consists of the X-10 wireless technology which allows communication between the remote control and a cart or between carts. Cart to cart communication consists of collision avoidance routines using a right-of-way scheme. The higher level communication is via a wireless RS-232 connection (less than 10 milliwatts) to the Master Controller which acts as a base for modem communication from CCRI headquarters. Don Nagy, the President of CCRI, routinely checks on the delivery history of CCRI robots by modem. The Master Controller also can direct the robot using the wireless LAN if complex paths are installed. The robot can be paused by pushing the stop button located on the front of the robot. However, once stopped, RoboCart has to be returned to its power source in order to be restated.
RoboCarts are working in a number of hospitals in the United States. The University of California, Los Angeles, has two carts which perform both intra and inter-laboratory delivery of medical specimens. Two robots avoid collisions while navigating on the same tape through the use of scheduling software. Special routines have been incorporated into these two robots which guide the robot back to the guide tape when they inadvertently become disoriented. A mobile robot leaves the guide tape for a variety of reasons including, dirt on the tape rendering it unreadable, stray light from an unexpected light source, and inadvertent or deliberate human intervention.
At The University of Virginia a single RoboCart is used to deliver microbiology and urine specimens to their respective analytical areas of the laboratory. These two analytical stations are at the extreme corners of a large core facility. Technologists save over 1.5 minutes for each trip completed by the robot. Since the RoboCart is operated exclusively within the confines of the laboratory, a plastic top is not employed.
Sourer Transporter (Sourer Automation Systems Inc., Holland, MI)
The Saurer Transporter, one of the many automated transport systems available from Saurer Automation, is larger than the RoboCart and yet smaller than the HelpMate (Figure 5). It has a longer footprint which allows it to carry a significantly higher payload. Like the HelpMate, it may have difficulty maneuvering in the tight confined of many laboratories. Therefore, the AGV may be used for bulk delivery of materials or specimens where large payload capacity is essential. Like the HelpMate, the Transporter is capable of riding on elevators.
Saurer Automation Systems (Holland, MI) offers a wide range of automated transfer devices to meet individual facility requirements.
Size, Guidance Method, Sensor Technology
The Saurer AGV comes in a wide variety of sizes and capabilities depending on the requirements of the transportation task. For example, the largest model of Transporter, the Tugger Integrator AGV, is from 6 to 7 feet long, 4.5 to 5.8 feet wide, and 2 to 5.25 feet high. The Tugger Integrator can carry payloads up to 6000 lbs at speed of 140 feet per minute. Their smaller product the Tote Integrator has dimensions of 5 to 7 feet in length, 2 to 4.5 feet in width, and 5.25 feet high. This product has a payload limit of 250 lbs at a speed of 110 feet per minute. The Saurer Transporter has an integrated control system that allows programmable dispatching using a computer or on-board keypad. For complex or changing paths, a PC based controlling system allows external management for routing and diagnostics. An onboard microprocessor is programmed to determine the shortest path to a given location. If multiple vehicles are using the same path, then the microprocessor calculates optimum routes to avoid collisions.
Guidance is provided by an invisible guide-path which is applied over tile, wood, carpeting, ceramics, and steel. The Ultraflex guide-path is manufactured by suspending fluorescent particles in an epoxy- like material. Almost invisible, Ultraflex will not wear out under normal use and cleaning for many months. The water based material can be applied by a trained individual in a few hours and dries quickly to a securely bonded film. Furthermore, normal amounts of abrasion will not diminish its ability to guide the Transporter. If the guide-path becomes scored by a gouge or scrape, the Transporter is programmed to ignore these obstacles. A break up to 5 inches can be tolerated without the need to reapply the film. This ability to ignore gaps facilitates using it over the threshold of elevators.
An ultraviolet light on the robot induces a visible emission from the guide-path allowing light sensitive diodes to detect it. Guidance is maintained within 0.3 inches even around corners. Stripes laid down perpendicular to the guide-path give instructions to the robot, for example, slow down or stop. Obstacle detection is provided by ultrasonic sensors backed up by contact bumpers.
Summary and Conclusions
Mobile robots are already providing labor savings to manufacturing facilities and over 60 hospitals nationwide. However, mobile robots are not the answer to all delivery needs. Each technology performs best in a distinct delivery domain depending on the capabilities of the guidance system, dimensions of the robot, and delivery speed.
Clearly, mobile robots will increase their utility when automatic pickup and delivery can be performed since this will obviate the need for continuous staffing at both ends of the delivery route. In the near future we can look forward to versatile, dependable, and affordable transportation in the form of intelligent mobile robots.