Medical Robotics and Computer-Integrated Surgery 52.3 Systems,Research Areas,and Applications 1209 Generally,TAB(...)is assumed to be a rigid transfor- One common class of methods is based on the iter- mation of the form ated closest-point algorithm of Besl and McKay [52.41]. TAB()=RABA十pAB, for example,3-D robot coordinates aj may be found for a collection of points known to be on the surface where RAB represents a rotation and PAB represents of an anatomical structure that can also be found in a translation,but nonrigid transformations are becoming a segmented 3-D image.Given an estimate T of the increasingly common.There are hundreds of methods transformation between image and robot coordinates, for computing TAB(...).The most common for medical the method iteratively finds corresponding points b;on robotics involve finding a set of corresponding geo- the surface that are closest to T&aj and then finds a new metric features TA and IB whose coordinates can be transformation determined in both coordinate systems and then finding a transformation that minimizes some distance function T+1=arg min >bj-aj）2. dAB distance [TB,TAB(TA)].Typical features can in- clude artificial fiducial objects(pins,implanted spheres, rods,etc.)or anatomical features such as point land- The process is repeated until some suitable termination Part marks,ridge curves,or surfaces. condition is reached. n 52.3 Systems,Research Areas,and Applications 52.3.1 Nonrobotic Computer-Assisted The main advantages of surgical navigation systems Surgery:Navigation are their versatility,their relative simplicity,and their and Image Overlay Devices ability to exploit the surgeon's natural dexterity and hap- tic sensitivity.They are readily combined with passive Medical robots are not ends in themselves.As the late fixtures and manipulation aids [52.43,44].The main Hap Paul often remarked,"the robot is a surgical tool drawbacks,compared to active robots,are those associ- designed to improve the efficacy of a procedure".(Dr ated with human limitations in accuracy,strength,ability Paul was the founder of Integrated Surgical Systems. to work in certain imaging environments,and dexterity Along with William Bargar,he was one of the first inside the patient's body (Table 52.2). people to recognize the potential of robots to fundamen- Because these advantages often outweigh the lim- tally improve the precision of orthopaedic surgery.)In itations,surgical navigation systems are achieving cases where the role of the robot is placing instruments widespread and increasing acceptance in such fields as on targets determined from medical images,surgical neurosurgery,otolaryngology,and orthopaedics.Since navigation is often a superior alternative.In surgical much of the technology of these systems is compati- navigation [52.42],the positions of instruments relative ble with surgical robots and since technical problems to the reference markers on the patient are tracked using such as registration are common among all these sys- specialized electromechanical,optical,electromagnetic, tems,we may expect to see a growing number of hybrid or sonic digitizers or by more general computer vi- applications combining medical robots and navigation. sion techniques.After the relationships between key coordinate systems (patient anatomy,images,surgical 52.3.2 Orthopaedic Systems tools,etc.)are determined through a registration process (Sect.52.2.6),a computer workstation provides graphi- Orthopaedic surgery represents a natural surgical cal feedback to the surgeon to assist in performing the CAD/CAM application,and both surgical navigation planned task,usually by displaying instrument posi-systems and medical robots have been applied to or- tions relative to medical images,as shown in Fig.52.8a.thopaedics.Bone is rigid and is easily imaged in CT and Although the registration is usually performed compu- intraoperative X-rays,and surgeons are accustomed to tationally,a simple mechanical alignment of an image doing at least some preplanning based on these images. display with an imaging device can be surprisingly ef- Geometric accuracy in executing surgical plans is very fective in some cases.One example [52.27]is shown important,for example,bones must be shaped accurately in Fig.52.8b. to ensure proper fit and positioning of components inMedical Robotics and Computer-Integrated Surgery 52.3 Systems, Research Areas, and Applications 1209 Generally, TAB(···) is assumed to be a rigid transfor￾mation of the form TAB vr A = RABvr A + pr AB , where RAB represents a rotation and pAB represents a translation, but nonrigid transformations are becoming increasingly common. There are hundreds of methods for computing TAB(···). The most common for medical robotics involve finding a set of corresponding geo￾metric features ΓA and ΓB whose coordinates can be determined in both coordinate systems and then finding a transformation that minimizes some distance function dAB = distance [ΓB, TAB(ΓA)]. Typical features can in￾clude artificial fiducial objects (pins, implanted spheres, rods, etc.) or anatomical features such as point land￾marks, ridge curves, or surfaces. One common class of methods is based on the iter￾ated closest-point algorithm of Besl and McKay [52.41], for example, 3-D robot coordinates aj may be found for a collection of points known to be on the surface of an anatomical structure that can also be found in a segmented 3-D image. Given an estimate Tk of the transformation between image and robot coordinates, the method iteratively finds corresponding points bj on the surface that are closest to Tk aj and then finds a new transformation Tk+1 = arg min T  j (bj − Taj) 2 . The process is repeated until some suitable termination condition is reached. 52.3 Systems, Research Areas, and Applications 52.3.1 Nonrobotic Computer-Assisted Surgery: Navigation and Image Overlay Devices Medical robots are not ends in themselves. As the late Hap Paul often remarked, “the robot is a surgical tool designed to improve the efficacy of a procedure”. (Dr Paul was the founder of Integrated Surgical Systems. Along with William Bargar, he was one of the first people to recognize the potential of robots to fundamen￾tally improve the precision of orthopaedic surgery.) In cases where the role of the robot is placing instruments on targets determined from medical images, surgical navigation is often a superior alternative. In surgical navigation [52.42], the positions of instruments relative to the reference markers on the patient are tracked using specialized electromechanical, optical, electromagnetic, or sonic digitizers or by more general computer vi￾sion techniques. After the relationships between key coordinate systems (patient anatomy, images, surgical tools, etc.) are determined through a registration process (Sect. 52.2.6), a computer workstation provides graphi￾cal feedback to the surgeon to assist in performing the planned task, usually by displaying instrument posi￾tions relative to medical images, as shown in Fig. 52.8a. Although the registration is usually performed compu￾tationally, a simple mechanical alignment of an image display with an imaging device can be surprisingly ef￾fective in some cases. One example [52.27] is shown in Fig. 52.8b. The main advantages of surgical navigation systems are their versatility, their relative simplicity, and their ability to exploit the surgeon’s natural dexterity and hap￾tic sensitivity. They are readily combined with passive fixtures and manipulation aids [52.43, 44]. The main drawbacks, compared to active robots, are those associ￾ated with human limitations in accuracy, strength, ability to work in certain imaging environments, and dexterity inside the patient’s body (Table 52.2). Because these advantages often outweigh the lim￾itations, surgical navigation systems are achieving widespread and increasing acceptance in such fields as neurosurgery, otolaryngology, and orthopaedics. Since much of the technology of these systems is compati￾ble with surgical robots and since technical problems such as registration are common among all these sys￾tems, we may expect to see a growing number of hybrid applications combining medical robots and navigation. 52.3.2 Orthopaedic Systems Orthopaedic surgery represents a natural surgical CAD/CAM application, and both surgical navigation systems and medical robots have been applied to or￾thopaedics. Bone is rigid and is easily imaged in CT and intraoperative X-rays, and surgeons are accustomed to doing at least some preplanning based on these images. Geometric accuracy in executing surgical plans is very important, for example, bones must be shaped accurately to ensure proper fit and positioning of components in Part F 52.3