MEMS-based systems can significantly improve accuracy in aligning hip and knee implants with a patient’s anatomy, reducing discomfort and the need for revision surgery.
Navigation is typically associated with cars, trucks, aircraft, ships, and, of course, people. It has also begun to play a significant role in medical technology, where it is used in precision surgical instruments and robotics. The design requirements of a surgical navigation tool share broad similarities with traditional vehicle navigation, but they also pose some distinct challenges—because the devices are used indoors, GPS assistance is not possible, for example—and they require a higher level of performance.
In this article, we will examine the unique challenges of medical navigation applications and explore possible solutions ranging from sensor mechanisms to system characteristics. Critical sensor specifications will be reviewed as well as potential error and drift mechanisms that should be taken into account during sensor selection. Enhancing sensors through integration, fusion and processing, by the use of Kalman filtering, for example, also will be highlighted. Before diving into the details, however, it may be useful to review some fundamental principles of inertial microelectromechanical systems (MEMS) sensor technology.
MEMS basics
Once considered exotic, MEMS is now a mature technology that most of us encounter on a daily basis. It makes our automobiles safe, enhances the usability of our mobile phones, measures and optimises the performance of tools and sports equipment and increasingly improves in- and outpatient healthcare.
 |
| Table I: Medical applications by motion type. |
MEMS devices used for linear motion sensing are typically based on a micromachined polysilicon surface structure built on top of a silicon wafer. The micromachined structure is suspended over the surface of the wafer by polysilicon springs, which provide a resistance against acceleration forces. Under acceleration, a deflection of the MEMS beam is measured by a differential capacitor made up of independent fixed plates as well as plates attached to the moving mass. Thus, movement unbalances the differential capacitor, resulting in a sensor output with amplitude proportional to acceleration. To use a familiar example, when an automobile is under sudden and excessive deceleration caused by a crash, the MEMS beams within the airbag sensor experience this same motion, which unbalances the capacitor and ultimately produces the signal that triggers airbag deployment. This same basic accelerometer structure, perhaps tuned to different application performance parameters and with additional data processing, can provide accurate indications of tilt, velocity and even position. A separate yet technologically related structure known as the gyroscope detects rotational rates, with outputs in degrees/second, as opposed to gs, or gravity, for accelerometers.
Converting motion detection into healthcare value
The ability to accurately detect and measure motion by means of a compact device that consumes minimal power can be valuable for nearly any application that involves movement, and even those where lack of motion is critical. Table I on the next page outlines some basic medical applications by motion type. More advanced applications presenting additional challenges will be discussed later.
Moving beyond simple motion sensing
While simple motion detection—linear movement along one axis, for example—can be valuable, most applications involve multiple types of motion occurring on multiple axes. Capturing this multidimensional state of motion not only enables new benefits, but it maintains accuracy in situations where off-axis disturbances may affect the measurement of a single primary axis of motion.
 |
| Figure 1: The six degrees of motion measurement required for full motion assessment: linear x, y and z motion, and roll, pitch and yaw rate rotation. |
In many cases it is necessary to combine multiple sensor types (linear and rotational, for instance) to precisely determine the motion an object has experienced. For example, an accelerometer can be used to determine an angle of inclination, since it is sensitive to the Earth’s gravity. In other words, as a MEMS accelerometer is rotated through a +/-1g field (+/-90o), it is able to translate that motion into a representation of an angle. However, the accelerometer cannot distinguish static acceleration (gravity) from dynamic acceleration. In this case, an accelerometer can be combined with a gyroscope, and additional data processing of the combined devices can discern the linear acceleration versus tilt (i.e., if the gyroscope’s output indicates a rotation coincident with the apparent tilt registered by the accelerometer). The sensor fusion process becomes more complex as the system dynamics (number of axes of motion and degrees of freedom of motion) increase.
It is also important to understand how the environment influences sensor accuracy. Temperature is an obvious concern, and can be corrected for; in fact, high-precision sensors are precalibrated and will dynamically compensate themselves. A less obvious factor to consider is the potential for even slight vibrations to produce shifts in accuracy of rotational rate sensors. These effects, known as the linear acceleration effect and vibration rectification, can be significant depending on the quality of the gyroscope. Sensor fusion, again, can play a role in improving performance by the use of an accelerometer to detect linear acceleration and applying this knowledge, along with a calibrated understanding of a gyroscope’s linear acceleration sensitivity, for correction.
Many applications require multiple degrees-of-freedom motion detection. For example, a six degree-of-freedom inertial sensor has the ability to detect linear acceleration on each x, y and z axis and rotational movement on the same three axes, also referred to as roll, pitch and yaw (Figure 1).
Navigation from vehicles to surgical instruments
 |
| Table II: Widely used navigational sensors, and their applicability to medical navigation. |
The industrial use of inertial sensors as a navigation aid has become widespread. Typically, they are used in conjunction with other navigation devices such as GPS. When GPS access is unreliable, inertial guidance can fill the coverage gap by using what is called dead reckoning. Except for the simplest navigation, most solutions will rely on multiple sensor types to deliver the required accuracy and performance under all conditions. GPS, optical and magnetic sensing are well understood and available. However, each technology has limitations, and taken together they cannot compensate completely for each other’s inaccuracies. MEMS inertial sensors have the potential to fully compensate for sensor inaccuracies since they are free from many of the same interferences and do not require an external infrastructure: no satellite, magnetic field or camera is needed—just inertia. The major navigational sensor technologies are outlined in Table II, along with their strengths and potential limitations.
Just as vehicle navigation devices are subject to GPS blockage, optical guidance technology used in medical systems can encounter line-of-sight blockages. Inertially-based sensors can perform dead reckoning during optical blockage incidents and enhance system reliability by providing redundant sensing.
Medical navigation
One medical application of the principles outlined in Table II involves the use of inertial sensors in the operating room to achieve a more accurate alignment of artificial knee or hip joints with a patient’s unique anatomical structure. The goal in this case is to align the implant with the patient’s natural axis to less than 1º error. Mechanical alignment, which is practised in more than 95% of total knee arthroplasty (TKA) procedures, typically results in 3º or greater error. Computer-assisted approaches using optical alignment have begun to replace some mechanical procedures, although uptake has been slow, probably because of equipment overhead. Whether mechanical or optical alignment is used, approximately 30% of these procedures result in misalignment (defined as >3º error), which leads to patient discomfort and often additional surgery. Reducing misalignment has the potential to reduce surgery time, enhance patient comfort and produce longer lasting joint replacements.
 |
| Figure 2: MEMS-based inertial measurement units provide six-degrees-of-motion measurement in compact form factors suitable for surgical instrumentation. |
Inertial sensors in the form of a full multi-axis inertial measurement unit (IMU) have been shown to provide substantial improvement in accuracy for TKA procedures. Devices such as the ADIS16334 (Figure 2) contain all of the required sensing—three linear and three rotational sensors—to replace mechanical- and optical-based alignment techniques. The device uses multiple sensor types and embedded processing to dynamically correct for sensor drift mechanisms, such as linear acceleration shift on gyroscopes, and temperature drift of linear and rotational sensing. A standard four-wire serial peripheral interface (SPI) provides a simple interface to this relatively complex suite of precision sensors.
MEMS inertial sensors are highly reliable (as witnessed by a 20-year track record in the automotive industry) and commercially attractive as demonstrated by their successful application in mobile phones and video games. However, performance levels vary significantly, and devices suitable for gaming will not be able to address the high-performance navigation problems outlined here. The key MEMS specifications of interest are bias drift, vibration influence, sensitivity and noise. Precision industrial and medical navigation typically require performance levels that are an order of magnitude higher than is available from MEMS sensors used in consumer devices. Table III outlines general system considerations that can help focus the sensor selection.
 |
| Table III: System goals/constraints help focus sensor selection. |
Most systems will integrate some form of Kalman filter to effectively merge multiple sensor types. The Kalman filter takes into account the system dynamics model, the relative sensor accuracies and other application-specific control inputs, and effectively makes the best determination of actual movement. High-accuracy inertial sensors (low noise, low drift and stability over temperature/time/vibration/supply variances) reduce the complexity of the Kalman filter, the number of redundant sensors required and the number of limitations on allowable system operation scenarios.
The complexity of medical MEMS
While a range of medical applications exist from relatively simple motion capture to complex motion analysis, the sector’s high-performance requirements pose both complex and computationally intense design challenges. Fortunately, many of the principles required for solving these next-generation medical challenges are based on proven approaches from industrial navigation problems including sensor fusion and processing techniques. Within medical navigation, the complexity of motion and requirements for precision and reliability will drive the need for multiple sensors, additional sensor postprocessing, sophisticated algorithms and complex test and compensation schemes.
In parallel with a strong push for small, low-power, multi-axis inertial sensors for consumer applications, there is an equally strong focus by some developers for environmentally robust, high-accuracy, low-power, high-performance sensors. These inertial MEMS devices offer advantages in terms of precision, size, power, redundancy and accessibility over existing measurement and sensing technologies.
Source:
