Abstract
A robot functioning in an environment may exhibit various forms of behavior emerge from the interaction with its environment through sense, control and plan activities. Hence, this paper introduces a behaviour selection based navigation and obstacle avoidance algorithm with effective method for adapting robotic behavior according to the environment conditions and the navigated terrain. The developed algorithm enable the robot to select the suitable behavior in real-time to avoid obstacles based on sensory information through visual and ultrasonic sensors utilizing the robot's ability to step over obstacles, and move between surfaces of different heights. In addition, it allows the robot to react in appropriate manner to the changing conditions either by fine-tuning of behaviors or by selecting different set of behaviors to increase the efficiency of the robot over time. The presented approach has been demonstrated on quadruped robot in several different experimental environments and the paper provides an analysis of its performance.
Introduction
Cancer killed 6.7 million people in 2002, with a projected figure of 10.3million by 2020. This occurs in women aged 40 to 79 and among men aged 60 to 79 (1). At the end of the 20th century more than 930,000 people died of cancer every year in 15 member countries of the European Union (EU) (2).
Recently, the fight against cancer has been targeted toward reducing mortality by converting a life-threatening disease to a chronic one, and even possibly achieving complete cure or response. Advances in molecular research and the knowledge of the aetiology promise targeted therapies tailored to inhibit carcinogenesis. These strategies for targeted cancer treatments (e.g. antiangiogenesis agents) avoid most of the side-effects related to general and cytotoxic treatments (3).
Surgery still remains one of the most effective treatments in the oncological field. Minimally invasive surgery has become the main objective in surgical research, to minimise the side effects of open surgery such as pain and blood loss. Laparoscopic cancer surgery started with liver biopsies that were the first laparoscopic procedures attempted by general surgeons in 1982 (4), four years later laparoscopy was applied to staging of pancreatic cancer with a reported accuracy rate of 93% (5). Surgeons first tried laparoscopic surgery on colon cancer in 1990, after it became popular for removing the appendix and gallbladder (6). Since then it has been established in many fields of oncological surgery, resulting in less pain, decreased infection and bleeding rates, and shorter hospital stay. There is a more rapid return to work and better cosmesis. Oncologically, laparoscopic surgery has demonstrated the same oncological results as open surgery in many procedures. For these reasons laparoscopy has become very popular among cancer surgeons and cancer sufferers (7).
Despite the advantages of laparoscopic surgery there remain barriers preventing its widespread adoption. Some of the laparoscopy limitations are the learning curve, the lack of propioception, spatio-temporal awareness and haptic feed-back, the compromise of hand-eye coordination, the restricted degree of movement of current laparoscopic instruments and the lack of correct ergonomics for surgeons involved in the procedure.
The restrictions of laparoscopic surgery meant in some oncological areas the open surgery maintained its advantages especially regarding oncological results and feasibility. In order to maintain the advantages of the minimally invasive approach but to avoid the restrictions of laparoscopic surgery, robotic surgery has emerged as a reliable surgical therapy that can achieve the same results of laparoscopic surgery, with a shorter learning curve. The surgeon has a more controlled degree of movement with an amplified three-dimensional visualization of the surgical field. Surgical robots have been designed to overcome laparoscopic surgery limitations and extend capabilities of surgeons. Robotic surgery facilitates the advantages of open surgery while refining the minimally invasive techniques [see Table 1].
Laparoscopic surgery vs. robot assisted surgery
Laparoscopic surgery vs. robot assisted surgery
Robotics in health is immersed in a global process of technological innovation since late the 20th century and early 21st century. Computer technology has provided a number of advanced surgical techniques that are effective tools for preoperative diagnostics and planning, intraoperative navigation, and robotic surgery (8).
The use of robotics in Medicine dates back 75 years, but only during the past 10 years has its potential been recognized around the world. In 1985 a robot, the PUMA 560 was used to place a needle for a brain biopsy under CT scan guidance. In 1988, the PROBOT was used to perform prostatic surgery (Institute of Urology, University College London). The ROBODOC from Integrated Surgical Systems was introduced in 1992 and is a robot to ream out precise fittings in the femur for hip replacement surgery. Furthermore, in the orthopaedic field the ACROBOT has been used for knee surgery developed at University College Hospital. Further development of master-slave robotic systems was carried out by Intuitive Surgical with the introduction of the da Vinci Robot and Computer Motion with the AESOP and ZEUS robotic surgical systems.
There are several available robotic systems and master-slave manipulators available to allow surgeons to perform minimally invasive surgery. These robotic systems have evolved from performing specific tasks during an operation, to becoming a virtual extension of the surgeon's hands, while providing unparalleled dexterity and views of the surgical field (Table 2).
Types of robotic surgical system and their application
Types of robotic surgical system and their application
In 1994, the AESOP 1000 system, developed by Computer Motion (Santa Barbara, CA, became the world's first surgical robot approved by the United States Food and Drug Administration. This was primarily a camera holding robot eliminating the need for additional operating personnel to perform that task. It was followed with AESOP 2000 in 1996, with the addition of voice control (HERMES). In 1998, the AESOP 3000 became available. In addition to camera control, it is an interactive surgical robot, with wristed instruments providing 7 degrees of freedom, giving the surgeon significant range of motion.
Aesop 3000 has 3 modes of movements: discontinuous, continuous, and pre-programmed. In the discontinuous mode small-step movements are activated with a single word indicating the direction required (i.e., “left”, “right”, “up”, “down”', “in”, “out”). In the continuous mode, the direction commands are preceded by the verb “move.” In the pre-programmed mode, three locations of the operating field can be saved in the software (9).
da Vinci™ Surgical System
The da Vinci Surgical System was developed by intuitive Surgical (Mountain View, CA) This system consists of three main parts, The Surgeon Console, which is controlled by the surgeon, the Surgical Cart, of which three to four arms are directly under the surgeon's control from the console are used to perform the procedure (10). The camera arm is also controlled from the console with the activation of a foot pedal. This camera has two endoscopes enclosed in a single sheath to allow a 3-dimensional image (3D). The master console consists of an image processing computer that generates the 3-dimensional image with depth of field. The surgical console is also equipped with foot pedals to control electrocautery, camera focus, and master control grips that drive the servant robotic arms at the patient's side (7) [FIGURE 1].
daVinci master slave robot draped for surgery. Note the height of the robot in the operating room
The Zeus telerobot was developed by Computer Motion, the manufacturer of AESOP®. The Zeus system provided the same function as the da Vinci except for the internal articulated endoscopic wrist. The right and left robotic arms replicate the arms of surgeon, and the third arm is AESOP voice-controlled robotic endoscope visualization. In the Zeus system the surgeon is seated upright with the video monitor and instrument handles positioned ergonomically to allow complete visualization of the operating room (7). Furthermore Zeus enabled surgeons to perform long-distance remote control surgery using SOCRATES® which is a surgical tele-collaboration system that links remote surgeons directly with other surgeons in the operating room (10). Unfortunately, the production of the ZEUS has ceased due to market forces.
Telesurgery
Telesurgery or remote surgery is the application of robotic surgery across long distances. It involves the precise integration of multimedia, telecommunications and robotic technologies.
In 2001 the Zeus system was used for a trans-Atlantic robotic surgery operation, between New York and Strasbourg. Connections between the sites were performed by high-speed terrestrial network, through asynchronous transfer mode (ATM) technology. The authors performed a robot-assisted laparoscopic cholecystectomy in 54 minutes. Despite the distance between surgeons and patients of more than 14000 km, the mean time lag for transmission during the procedure was only 155 msec (11).
Remote surgery depends on robotic assistance and information technologies. Telesurgery potentially gives surgeon the ability to mentor each other from remote locations. It makes training and education more accessible, and may improve the standard of surgical care throughout the world (11).
In fact telementoring has been reported in three craniotomies for brain tumours, a craniotomy for an arteriovenous malformation, a carotid endarterectomy and a lumbar laminectomy. A robotic telecollaboration sytem (Socrates; Computer Motion Inc., Santa Barbara CA) was used in all cases. Four integrated services digital network (ISDN) lines with a total speed of 512 kilobytes per second provided telecommunications between a large academic centre in Nova Scotia and a community-based centre 400 Kms away. All the procedures were performed without complication and it was concluded that long-distance robotic-assisted telementoring, although still in its infancy, is feasible, reliable and safe (12).
Applicability of robotic technology in oncological surgery
Oncological surgery involves large open incisions (laparotomy, thoracostomy, lumbotomy). There is an associated blood loss, longer recovery period and painful postoperative period that requires more analgesia. Laparoscopic surgery avoids most of the side effects of open surgery while maintaining oncological integrity. Furthermore robotic surgery maintains the advantages of laparoscopic surgery but reduces some of its limitations such as learning curve.
Robotic oncological surgery is a targeted therapy, focused on tumour treatment avoiding the general side-effects of open surgery. The recent development of robotics in surgery has demonstrated that it is a feasible and reliable therapeutic approach for cancers (Table 3).
Procedures in robotic surgery
Procedures in robotic surgery
Laparoscopic surgery for colorectal cancer has been shown to be safe and feasible, but highly challenging, with a steep learning curve compared to a simpler procedure (laparoscopic cholecystectomy). Suturing for colonic resection has a high difficulty rate, and has been overcome by stapling devices and extracorporeal surgery. The Robotic application for colorectal surgery allows for a superior 3D imaging with magnification and improved dexterity achieved with the endowristed robotic instruments. Recently a series of 30 consecutive colectomies was published. Six were surgeries for cancer, 16 for polyps and 8 for diverticulitis. A right colectomy (n=17) took longer than a sigmoid colectomy (n=13) because the anastomosis in the former were all performed intracorporeally. Two patients needed conversion to open surgery. The authors concluded that the cost of the device, the increased room requirements to accommodate the equipment and the difficult team communication with the surgeon in the console, were the major disadvantages. All these issues become less of a problem as the team passes its learning curve for not only the surgery, but also the setup of the robot. The authors also reported disadvantages regarding the inconvenience of rotating the patients during the procedure to allow gravity to retract organs, or the inconvenience of altering port placement of the camera during the case. These disadvantages are more related to the planning of the surgery. Robotic surgery uses laparoscopic access to the surgical field, but then the movements and skills required become dissimilar. For example, the port and patient position are crucial to allow adequate robotic arm movement. The authors remark that the advantages of robot assistance for colorectal procedures were the enhanced vision, the surgeon-control of the camera, the 7 degrees of freedom of movement for the endowrist instruments and the reduced surgeon fatigue (13). The present data demonstrated that robotic-assisted colorectal surgery is feasible and safe, and that improves some of the limitations of the current laparoscopic approaches.
Gynaecological malignancies
Robotics has been utilized mainly for the treatment of three benign gynaecologic conditions: benign hysterectomy; repair of vesicovaginal fistula; and sacrocolpopexy. In these conditions robotic assistance has demonstrated that it is a feasible, reliable and safe (14;15). In regard to gynaecological malignancies the use of the robotic approach is still in development. Nevertheless taking into account that laparoscopic-assisted vaginal hysterectomy is the technique of choice even in adenocarcinoma of the uterus it is reasonable to expect that the robotic approach would be adopted in gynaeoncological surgeries. The magnified vision and the improvement in degrees of movement could facilitate the procedure. Recently it has been reported in a series 30 robot-assisted hysterectomies, that 12 of them were performed for malignant disease stage I. The mean operative time was 185 min (range 43–315), the mean blood loss was 83 ml (range 0–900) and the mean stay of patients was 8 days. The rate of complications was 17%. The study highlighted the advantage of using robotic-assistance within a confined space (extended hysterectomy with pelvic lymph-node dissection; (16).
Neurosurgical tumours
Neurosurgeons were pioneers of robotic surgery. The concept of “brain surgery without incision” became a reality with the approval of gamma knife in 1982 (10). The PUMA 760 robotic arm was the first robot used in neurosurgery. They used programs which calculated the
CyberKnife has also been applied to the treatment of spinal lesions. This was a prospective cohort evaluation of spinal radiosurgery in 125 spinal lesions in 115 consecutive patients. All of them were treated with a single-fraction radiosurgery technique. There were 17 benign tumours and 108 metastatic lesions. Tumour volume ranged from 0.3 to 232 cm3 (mean, 27.8 cm3). In this large prospective evaluation the CyberKnife system was found to be feasible, safe, and effective. This technique of radiosurgical ablation of spinal lesions was advantageous due to short treatment times in an outpatient setting with rapid recovery and symptomatic response. The study has shown that this technique may be applied to a variety of spinal lesions as a primary treatment or for lesions not suitable to be submitted to open surgical techniques, in medically inoperable patients, in lesions located in previously irradiated sites, or as an adjunct to surgery (20).
Thoracic
A minimally invasive approach for major thoracic procedures has been limited because of two-dimensional vision and limited manoeuvrability of the instruments in the small surgical field. Robotic devices with three-dimensional imaging and excellent freedom of movement would allow the expansion of the minimally invasive treatments for thoracic malignancies (21).
Recently, a series of 30 patients with non-small cell lung cancer stage I disease except two (carcinoid tumour and B-cell lymphoma) were submitted to da Vinci robotic video-assisted thoracic surgical lobectomy (VATS). They reported no operative mortality, but a 12% conversion to open surgery. Once again the study suggested that Robotic assistance for dissection, isolation and division of pulmonary hilar structures is feasible and safety. Interestingly they suggested that the robotic approach offered advantages over the laparoscopic technique because of a superior, stable image for the surgeon and because of the endowrist robotic instruments that allowed more accurate hilar dissection. The oncological results were promising. (21).
Laparoscopic thoraco-oesophageal surgery may have better survival outcomes as a more accurate lymph node dissection can be performed. The application of robotic thoraco-oesophageal surgery is relatively novel. Recently a series of 6 robotic oesophagectomies, where four were cancers and had lymph node dissection, was published. They reported only one case of persistent lymph fistula and all the procedures were completed by robot assistance. Clearly, the oesophagus is a major organ of interest for robotic surgeons with advantages of the robotic approach shown in this study (22).
Urological tumours
Robotic Uro-oncology is an expanding field, particularly as an option for the treatment of localized prostate cancer. Laparoscopic radical prostatectomy is a challenging procedure for all the reasons stated previously. The daVinci robot with three-dimensional visualisation is helpful for preserving bladder neck and neurovascular bundles, and at the same time the endowrist allows simplification when performing of the urethral anastomosis. In our experience robot-assisted radical prostatectomy (RAP) is technically demanding but feasible, with the patient clearly benefiting. We recently reported our experience with daVinci(R) robot assisted extraperitoneal laparoscopic radical prostatectomy. A total of 325 patients underwent RAP for clinically localized prostate cancer at our center during a 2-year period. Average total operative time was 130 minutes (range 80 to 480). Intraoperative data included a mean blood loss of 196 cc with no open conversions. Bilateral, unilateral and nonnerve sparing prostatectomy was performed in 70%, 24% and 6% of patients, respectively. Of the patients 96% were discharged home within 8 to 23 hours of surgery. In our experience the extraperitoneal approach offers the advantages of improved dexterity and visualization of the robot, while avoiding the abdominal cavity and potential associated morbidity. As surgeons gain more experience with this new technology, the extraperitoneal approach simulating the standard open retropubic technique is likely to gain popularity (23). On the other hand we had previously compared our results in laparoscopic radical prostatectomy (LRP) and RAP. The two techniques were compared in 100 patients with localised prostate cancer who had LRP or RAP (n=50 in each group). Functional and oncological results were similar in both groups and interestingly the blood loss was less in the RAP group, with no blood transfusions in either group. This difference could be related to the superior magnified, three-dimensional image of the robot that allows more accurate haemostasis (24).
Other authors had performed the same procedure using a transabdominal approach. In 2003 a single-institution, prospective, unrandomized comparison of histopathological and functional outcomes, at baseline and during and after surgery, in 100 patients undergoing retropubic prostatectomy (RRP) and 200 undergoing robotic transabdominal prostatectomy was performed. Most robotic prostatectomies (93%) and none of the RRP patients were discharged within 24 hours; the duration of catheterization was twice as long after RRP (15.8 vs 7 days). The robotic procedure appeared to be safer, less bloody and required shorter hospitalization and catheterization. The oncological and functional results were favourable in patients undergoing robotic prostatectomy (25). Recently the group has published the results on 154 consecutive men submitted to robotic prostatectomy with nerve preservation. At 1 year of follow-up 96% of the men reported having had intercourse and 71% had recovered normal erectile function. As well, 97% of the men were continent. In regard on oncological results, among the patients with organ-confined disease, 4.6% has positive surgical margins and no patient had PSA recurrence at 12 months (26).
The robotic surgery modality is also being applied to renal cancer, bladder cancer and retroperitoneal lymph node dissection in testicular cancer.
In renal cancer laparoscopic surgery has been established as the gold standard treatment in renal masses less than 7 centimetres. In a comparison series between laparoscopic and open surgery for T1/T2 N0M0 renal cancer, after a median follow-up of 73 months, laparoscopic radical nephrectomy for renal cancer had demonstrated that is oncologically equivalent to open radical nephrectomy with the same 5 and 10-year cancer specific and actuarial survival rates (27). Taking into account that laparoscopic surgery is being established in renal cell cancer treatment the next step is the consolidation of robotic surgery in this field. There had been reported initial series in the application of robots in radical and partial nephrectomy (28;29). Both techniques had been proved as feasible and reproducible.
The application of laparoscopy and even more of the robotic surgery to the treatment of bladder cancer is still in its infancy. Radical cystoprostatectomy is the gold standard surgical treatment or invasive bladder cancer but despite this it is associated with a significant comorbidity. It is reasonable to expect that laparoscopy and robotic surgery could minimise the rate of complications as it has been demonstrated in other surgeries. Initial results with robotic radical cystectomy have demonstrated that is feasible and associated with low comorbidity. The urinary diversion could be performed extra or intracorporeally although performing the whole procedure intracorporeally is associated with a longer operation time (30;31). Recently a comparative analysis of early outcomes of open radical cystectomy and robotic-assisted cystectomy was shown. The study cohort comprised 37 consecutive patients undergoing radical cystectomy; 24 cases were performed by the conventional open method and 13 by the robotic method. The robotic method resulted in statistically significant lower median estimated blood loss, shorter hospital stay, and longer operating time compared with the open group (32).
Nevertheless it has been purposed that indications for robotic-assisted surgery in bladder cancer must excluded those patients with local advanced disease, as well as those with neoadjuvant chemotherapy in order to achieve better results (33), in our centre we are developing a new program for laparoscopic treatment of bladder cancer. We perform laparoscopic radical cystectomy routinely and we have opened the program to performing anterior and total exenterations in patients with locally advanced bladder cancer. Our initial experience allows us to conclude that these laparoscopic procedures are feasible, reliable and safe and now that the laparoscopic schedule is established we expect that they could be performed robot assisted to allow other surgeons to learn and perform this demanding surgery.
Current status and future directions of robotic oncological surgery
While, in 1976 there were only a few hundred industrial robots, by the end of 1999, there were over one million. At the beginning of 2000, there were approximately 350 robots employed actively in the surgical field worldwide. This corresponds to a growth rate of 20%, which is analogous to that seen in the industrial area (34). This data only considers surgical systems such as da Vinci and Zeus, but does not take into account other smaller robots all over the world.
There are however a number of disadvantages which prevent their widespread applications. The cost of a robotic surgical system is a million euro with an annual maintenance of 150000€. The high costs of these systems prohibit their use at many centres. It is possible that cost reduction may occur with time, and technological improvement. However the research necessary for the development of these systems may itself fuel further price increases (7). The novelty of robot assisted surgery and the lack of long term data showing efficacy and efficiency similar to open surgery complicate third party reimbursement for these procedures, making them less affordable. (8). However, as robots become more popular and the market has more providers the cost will ultimately come down, as with computers and mobile phones.
Currently robotic surgical systems consist of cumbersome robotic arms and surgeon's console. These robots are complex and require a large operating room. Hospitals incur the cost of restructuring the operating room, to enlarge the room, and equip it with the necessary hard and software. This contributes to high initial costs for hospitals embarking on robotic surgery programs. Miniaturization technology promises to solve these problems associated with bulky present day robots. As with the computer and mobile phone technology happened the surgical devices will become smaller and cheaper. In addition, the other problem is robots deprive surgeons of their sense of touch (haptic feedback). They have improved the laparoscopy limitations with regard to the learning curve, by using three-dimensional vision which helps propioception and spatio-temporal awareness. Recently, studies in this area have led to the development of two types of sensorial systems: force-feedback systems that allows the determination of the force that is being applied to the tissues, and vibration-tactile systems that allow the discrimination of different tissue consistencies (34).
Currently robotic surgery has a role in oncological surgery as a minimally invasive technique, avoiding the general side-effects associated with open surgery and minimising the challenge of some oncological laparoscopic procedures. Although further studies may be warranted to establish the oncological results of robotic surgery in cancer, the present data allows the potential establishment of robotic programs in most of the oncological specialities. While these technologies mentioned above are being established, the next generation of robots are already on the horizon.
Future robots
With advances in microelectromechanical systems, microrobots and eventually nanorobots are quickly becoming a reality. We recently reported on the use of microrobots (Figure 2) to assist in urological surgery. (35) Other reports have also been published on the use of these devices to assist in laparoscopic cholecystectomy. (36) Microrobotic cameras placed inside the abdomen were used to carry out surgical procedures. These robots provide the surgeon with a 360 degree view of the abdominal cavity. The views provided by these miniature camera robots give the surgeon an added frame of reference. This improves the surgeons spatio-temporal orientation and overall precision in carrying out delicate tasks laparoscopically. One advantage of these robots is that they allow the surgeon to retain haptic feedback provided through the laparoscopic instruments. These systems unfortunately were tethered for power, although wireless systems are available they can only be used for short procedures due to poor battery life. Another challenge faced by these systems is the lack of an even surface inside the abdomen to facilitate navigation. These microrobots can easily be caught in the loops of bowel, which requires the surgeon to use the laparoscopic instruments to position them at specific sites in the abdomen.
Roving microrobot measuring 4cms across with a central camera
These microrobots although promising have yet to be used during human surgery. In their current state offering only an additional view of the surgical field, their usefulness may be quite limited. With the addition of effecter arms capable of carrying out specific tasks during the surgery, or sensors to monitor the surgical environment, their functionality will improve adding significantly to their appeal and possible adoption. Increased miniaturization has occurred at a nano level, hence nanorobots. A number of experiments have been reported where robots measuring 670 microns tall, and 170–240 microns wide have been used to move objects not visible with the naked eye (37) These nanorobots can be made mobile and are capable of working submerged in liquids such as urine and blood, carrying out specific tasks. They can potentially serve to diagnose and treat at the cellular level.
These emerging micro and nanorobots offers great promises, but also have significant challenges. Improved functionality, miniaturization and refinement of these robots, along with improvement in energy storage are paramount for micro/nanorobotic surgical tools to gain a foothold in the surgical arena.