FIELD OF INVENTION
This invention relates to an automated way of customizing and optimizing the fit of a bicycle (or other human powered conveyance) to a particular person.
Historically, most individuals who purchase a bicycle were steered toward a bicycle frame that appeared to be a good fit for their height and stature. Over time, many bicycle shop employees have become experienced at approximating what frame size will be a reasonable match. A bicycle can be chosen from off-the-shelf sizes that range from 13-26 inches. This is a measurement of the distance from the center of the pedal crank bearing (bottom bracket) at the bottom of the bicycle to top of the top tube that connects the saddle to the handlebars. Once a bicycle was chosen, the only refinements to the selected bike were typically limited to minor adjustments made to the height of the saddle.
More recently, two options for bicycle fitting have become available. The first of these is a fully manual technique, which measures basic body dimensions, including inseam and arm length. These dimensions are usually taken using a tape measure, and then these dimensions, along with subjective knowledge of the art of bicycle fitting, translate into either a selected frame size, adjustments to the components on the frame, or both. This approach is a static fit, and yields an approximate adaptation of the body size to the bicycle. What follows typically is a brief test ride by the buyer, who very subjectively evaluates the quality of the fit (generally without any knowledge of what constitutes a good fit).
The second, more recent, option involves performing a dynamic bicycle fit and uses an optical tracking system and a computer. This analysis utilizes either a bicycle that the cyclist already owns, one that the cyclist may purchase if the fit is successful, or a “bicycle simulator” with an adjustable frame. In the case of a traditional bicycle, the rear wheel is held off the ground with a mechanical stand that permits pedaling of the bicycle at a stationary location. (The stand may also include a separate power output sensor and indicator.) This test is dynamic, with the cyclist pedaling the bicycle while the cyclist is videotaped from various angles. The computer may convert the optical information into a stick figure. The information in such graphical pictures is analyzed by the system operator for the quality of the fit to the individual or for adjustments to components on the bicycle or the cyclist. Automated analysis by the computer may include certain generic, automated distance and angle measurements. Such a system is described in Andy Pruitt's Medical Guide for Cyclists (available on the Internet as an e-book from www.roadbikerider.com). It employs a Motus marker tracking system (www.peakperform.com, Vicor Peak, Colorado Springs, Colo.), which is a generic motion tracking system, not specifically designed for bicycle fitting.
- BRIEF SUMMARY OF THE INVENTION
Andy Pruitt, a leading expert in the field, explains in his aforementioned book why dynamic fitting is preferable. However, his methods and his system require experiential intuition and are at best only partly quantitative and automated. The optical tracking system is generic and not specific to bicycle fitting. What is lacking is a system (a) which is more quantitative and automated, (b) which is specific to bicycle fitting, (c) which leverages a computerized quantitative database of experience, and (d) which is easy for a less qualified bicycle shop employee or racing team mechanic to use.
The purpose of the present invention is to perform a semi-automated, low cost, high accuracy bicycle fit for an individual. One objective of the invention is to determine the proper bicycle frame size and geometry for the individual who is purchasing a new bicycle, possibly from a database of geometries of available models. A second objective of the invention is to determine how to adjust a bicycle so that the bicycle-to-cyclist interface is optimized. A third objective of the system is to determine which components on a bicycle could be changed so that the bicycle-to-cyclist interface is optimized. A fourth objective of the system is to determine which components should be added to the bicycle or the cyclist, or deleted from the bicycle or the cyclist, so that the bicycle-to-cyclist interface is optimized. This optimization can be directed at riding comfort, riding endurance, riding performance, or some other target criterion.
The invention is a system that performs precision measurements of body motions while a rider (cyclist) pedals a bicycle, potentially with realistic resistance, which itself might be used to quantify power output. The invention tracks the repetitive motions of pedaling a bicycle, and provides both visual outputs and quantitative measurements of the repetitive motions, which are then analyzed for correctable variations or anomalies in the movements. The analysis is dynamic, integrated, and synchronized.
In use, the invention is operated by “harnessing” the individual to be fitted. This process consists of attaching sensible markers—such as light emitting diodes (LED's)—to the individual's foot, ankle, knee, hip, shoulder, and wrist or other such locations. These markers are attached by means of a double sided adhesive pad, Velcro® straps, elastic straps, or similar fasteners. The markers are attached such that they can be viewed from the left or right side of the individual or both. A marker location tracking system (optical tracker) capable of detecting the moment-by-moment 3-dimensional locations of the markers (as well as the times of their acquisition) is positioned in proximity to the stationary bicycle, such as to one side. The individual is then asked to pedal the stationary bicycle while the tracker captures data. The test is then repeated on the other side of the individual's body. Alternatively, if the system employs a more sophisticated tracker (or multiple trackers), the data from more than one viewpoint (sides or front) could be acquired simultaneously.
A preferred method of use of the present invention gathers dynamic data to determine the optimum size of an “in-stock” bicycle for an individual purchasing a new bicycle. For this use of the invention, manual adjustments could be made to a commercially available bicycle between repeated sessions using the present invention. For example, the height, angle, and setback (fore-aft) position of the saddle are typical adjustments. Similarly, the height and angle of the handlebar may be adjustable. In addition, after-market components could be evaluated based on their proper fit to the cyclist being fitted. Such components include, but are not limited to, handlebar, saddle, pedals, and shoes. For example, the handlebar can be positioned forward or upward by replacing the handlebar stem with a longer one. Additionally, the angle of the handlebar within the stem can be adjusted. Similarly, the crank arms of the pedals may be replaced to change their length as necessary, or spacers or wedges may be added to the pedals or shoes.
In another preferred method of use of the invention, the data gathered by the system is used to determine the dimensions for a custom frame to be manufactured to fit the particular cyclist being fitted. The dimensions obtained from measuring the individual are used in a manual or automatic manner to manufacture the frame built to that cyclist's specifications. As with other methods of use, after-market or custom components could be evaluated for attachment to the custom bicycle frame. An advanced form of the invention could even suggest components from a database of component options of known dimensions and geometry.
In an additional method of use of the invention, the data gathered by the system is used to perform manual adjustments or changes to a bicycle that a cyclist already owns. In this embodiment, the data obtained from the invention is used to analyze opportunities to adjust the bicycle or the riding position favorably to a more comfortable or more efficient riding position. The data can also be used to determine opportunities to employ substitute bicycle components to provide a more comfortable or more efficient riding position. Upon completion of adjustments or component substitutions, the test is repeated. An updated set of data is gathered, and compared to the previous results for changes. This process applies to all adjustable or alterable dimensions of the bicycle and cyclist riding position. The process may be repeated as many times as is necessary or desirable to optimize all of the adjustable geometries available to the cyclist.
In an advanced embodiment of the apparatus of the invention, the data gathered by the system is used to dynamically and automatically adjust a bicycle simulator to the optimum riding position for the cyclist. In this embodiment, information from the system is fed back on a real-time basis to adjust various individual bicycle fit parameters sequentially and dynamically in a way that optimally modifies all dimensions that are critical to a proper fit. These changes are made on a real-time, automated basis while the cyclist is riding a bicycle fitting apparatus (simulator). This advanced embodiment of the present invention could employ computer-controlled motors, pneumatics, or hydraulics to modify the simulator's geometry automatically and dynamically to fit the cyclist. The output of the session would be the set of dimensions to which to adjust an existing or new bicycle in order to achieve the same “fit” on that bicycle as was obtained on the bicycle simulator. Alternatively, the dimensions would serve as specifications for a custom built bicycle.
DESCRIPTION OF THE DRAWINGS
In another embodiment of the invention, a database is maintained in which are collected the bicycle fit measurements of many successful fittings, which have been performed previously. This data is used to aid in future products and services as may be developed. This data may also be reused when a previously fitted individual purchases a new bicycle or wishes to have the fit optimized to a different choice of comfort, endurance, or performance.
The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate a preferred embodiment of the present invention and, together with the description, serve to explain the principle of the invention.
FIG. 1 is a simplified perspective view of the major components of this invention as an apparatus.
FIG. 2 is a flow chart summarizing the method of performing a bicycle fit using the invention.
FIG. 3 is a flow chart summarizing the automated portion of the method of performing a bicycle fit using the invention.
FIG. 4 depicts an example of what might appear on the screen of the computer operating the present invention.
- DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
The figures contain the following numerically identified elements:
- 1 a cyclist for whom a bicycle fitting session is being performed
- 10 a bicycle, training bike, bicycle fit simulator, exercise bike, or the like
- 11 a bicycle support stand, trainer base, or similar support
- 12 a bicycle frame (optionally supplied by the cyclist)
- 13 a bicycle handlebar (optionally supplied by the cyclist)
- 14 a bicycle saddle (optionally supplied by the cyclist)
- 15 each of two pedals (optionally supplied by the cyclist)
- 16 each of two wheels (optionally supplied by the cyclist)
- 20 an implicit 3-dimensional coordinate system
- 22 a wiring harness for powering and controlling markers 24
- 24 each of a plurality of trackable markers (such as light emitting diodes)
- 28 at least one marker tracker able to locate each marker in 3-dimensions
- 30 a computer system (such as a laptop) and operating system
- 32 a data path between each tracker and the computer system
- 34 application software (instructions and data) running on computer system 30
- 40 optional power output sensor
A preferred embodiment of this invention, as an apparatus, is shown in FIG. 1. FIG. 1 includes a cyclist 1, which of course is not part of the invention apparatus proper, and it shows a conventional bicycle 10. The bicycle comprises a frame 12, handlebar 13, saddle 14, each of two pedals 15, and (generally) each of two wheels 16. Because the bicycle must be pedaled in a stationary location, a support stand 11 is normally required. The support stand 11 should allow the rear wheel to rotate but hold the bicycle in a stationary and level position. The bicycle may be one owned and supplied by the cyclist, may be a potential purchase by the cyclist, may be an exercise bike, or may be an adjustable trainer or simulator (such as one designed specifically for fitting purposes). In the latter two cases, there may be less than two wheels 16 and the support 11 may be built in, but there will be some sort of frame 12, handlebar 13, saddle 14, and two pedals 15 which presumably are offset from and rotate around a rotational axis in the usual way.
For simplicity of explanation but without limiting the generality of the invention, we will assume a cyclist and a conventional bicycle hereafter unless otherwise noted. Nevertheless, the frame 12, handlebar 13, saddle 14, and pedals 15 could instead be those of a tricycle (with three wheels 16), a recumbent bicycle (with a chair-like seat instead of a conventional saddle 14), a unicycle, a motorbike (with pedals 15), a rickshaw, an exercise machine, a rowing bike, a human-powered water vessel, or other human-powered device, for example. However, because of sheer numbers, the most important application of the invention is to conventional bicycles. Nevertheless, in spite of the bicycle-oriented description of a preferred embodiment, the claims below will use more general terms.
FIG. 1 also shows wired LEDs used as trackable markers 24 affixed to appropriate locations on the body of the cyclist 1. (Desirable locations for the markers will be suggested later.) Because the markers 24 are actively illuminated in this preferred embodiment, they are powered and actuated electrically through a wiring harness 24. An alternative implementation might use optical fibers instead of electrical wires. Furthermore, the harness 24 might include batteries and a radio or infrared receiver so that there is no electrical cable between such a harness 24 and the computer 30 which otherwise powers and controls the LED activation.
The markers 24 are attached to the cyclist 1 or the cyclist's clothing so as to minimize movement with respect to the body part of the cyclist to which it is attached. The attachment means could be double-sided adhesive tape or foam, Velcro® patches or strips, elastic bands, or the like. The major requirements are that each marker 24 remain essentially immobile with respect to the closest bone or joint of the body and maximize visibility of the markers 24 by the marker tracker 28.
Preferred locations for the markers 24 are the following (and these particular locations will be assumed subsequently in the description). Each is associated with their brief descriptive nickname: The “foot” marker is affixed on the side of the cycling shoe vertically above the pedal axis and farthest from the frame. The “ankle” marker is affixed to the outside (farthest from the frame) of the ankle directly on the most prominent boney protrusion at the bottom of the fibula. The “knee” marker is affixed directly on the lateral protrusion of the upper fibula, and preferably not on the patella (knee cap). The “hip” marker is affixed on the boney prominence of the crest of the ilium (pelvis). The “shoulder” marker is affixed where the upper end of the humorous (arm) bone extends most outwardly when the arm hangs down. (This is the least obvious landmark location) The “wrist” marker is affixed on the lateral boney prominence of the wrist (the distal prominence of the radius arm bone). These landmark locations are preferred, because they (a) are most rigidly related to the skeleton of the cyclist, (b) should be rather unambiguous to find on most cyclists 1, and (c) are favorable for nearly exact repeatable placement at some future time. Clearly, additional or alternative locations (such as the elbow) could be used.
The moment-by-moment three-dimensional (3-d) location of each individual marker 24 is determined in real time by tracker 28 multiple times per second. An example of such a tracker 28 is the FlashPoint 5500 system manufactured by Boulder Innovation Group (www.boulderinnovators.com, Boulder, Colo.). It flashes each LED marker sequentially for a few milliseconds and uses multiple charge-coupled devices (CCDs) in the tracker 28 to determine the 3-d location of each LED and computes its coordinates in millimeters with respect to a 3-d coordinate system 20. These coordinates along with a millisecond-accuracy time stamp and a status code in a digital format are provided through a serial data cable 32 to application software running on computer system 30. The status code, for example, would indicate whether the LED was visible at the time of attempted acquisition.
The markers 24 and marker tracker 28 are herein assumed to be part of an actively pulsed optical tracking system. Alternative marker tracking systems could be used instead, such as those which use conventional video cameras and passive markers (typically retro-reflective spheres 1 centimeter or more in diameter). An example is the Motus system from Vicon Peak (www.peakperform.com, Colorado Springs, Colo.). However, passive systems operate less reliably for several reasons: (a) reflective markers which are smudged or partially eclipsed yield inaccurate location centroids; (b) the markers cannot be unambiguously identified especially when the trajectories of two cross; and (c) specular reflections off shiny surfaces in the field of view cause anomalous, extraneous marker detections. Generally such anomalies require manual, interactive post-processing. Therefore, a tracker 28 which tracks active sequentially illuminated markers 24 is described as part of the preferred embodiment of the present invention.
Magnetic-based or sonic-based tracking systems have further disadvantages (such as distortion around metallic, conductive objects or limited acquisition speed respectively) and are even less practical for the present bicycle fitting invention, but presumably could be used instead of an optical system.
Computer system 30 preferably is a contemporary desktop or laptop personal computer running a conventional operating system such as Windows XP (Microsoft Corp.), MAC OS X (Apple Computer), or Linux. Alternatively, system 30 could be a custom, proprietary combination of computing hardware, a liquid-crystal display (LCD) screen, and software. Computer 30 is programmed with application software 34, which directs operation of the markers 22 and tracker(s) 28 and acquires marker location data therefrom. Details of the procedural steps of that software are given in the flowchart of FIG. 3. An example of the graphical output of the software is shown in FIG. 4.
During operation of the invention, the locations of the markers 24 on the cyclist's body 1 are tracked by the system in real time and saved as XYZ coordinates with respect to a three- or four-dimensional coordinate system (three spatial dimensions, possibly including time too). If additional markers 24 are also attached to the bicycle 10, these would be tracked as well to monitor lateral sway of the bicycle 10 and to relate the cyclist's movements relative to the instantaneous position of the frame 11. Such additional markers could be used to define an XYZ coordinate system which is relative to the bicycle rather than to the tracker itself. Each marker 24 is tracked for its location every few milliseconds. Optionally, but preferably, the location coordinates are augmented with time coordinates (timestamps) indicating when each marker location was sensed (for example, in milliseconds since the beginning of the current system session). This allows computation of pedaling speed and better correlation between marker locations at the same moment in time. At least one data file is created for the sequence of coordinates of the markers' spatial locations and corresponding timestamps, and each marker's spatial and time coordinate datum is associated with the marker to which it corresponds. The data is preferably preserved for future reference. This same raw coordinate data is imported into programmed computer software 34 for analysis—either immediately or later or both. The software 34 may be a programmed spreadsheet (preferably with graphical plots as well as numeric computations) or may be custom software written in a programming language such as C++, for example.
Data can be acquired for a duration of just a few seconds or over a period of several minutes. Taking data over a period of several minutes affords the ability to look for trends, changes as the cyclist warms up, or unexpected shifts in body position on the bicycle. Furthermore, if the subject's output power is periodically measured by optional power output sensor 40 and the measurements correlated to the location and time data, further analysis can also quantify how cyclist position and bicycle configuration relates to performance or efficiency.
An advanced embodiment would further measure the cyclist's effort by monitoring one or more of the following physiological functions: heart rate, respiration rate, and blood oxygen level. For example, the moment-by-moment data from a commercial pulse oximeter could also be supplied electronically to the computer software 34 for correlation with the power output to calculate the cyclist's efficiency. Then the computer software 34 would use that to rate the bicycle configuration or cyclist's posture.
Once imported into the analysis software 34, the trajectories of each of the tracked markers are analyzed for anomalies, statistical characteristics (such as means and extrema), and transitions outside of expected limits. Suspect data might be discarded, such as those with anomalous, sudden changes in speed, for example. The limits are obtained from a database, which is representative of many individuals' data sets, which are considered to represent a good fit for the target criterion. Furthermore, the center and radius of the ellipse which most closely fits the trajectory of the “foot” marker might be computed, as well as statistics of the deviation of that trajectory from the ellipse. In addition, the cyclical motions of other markers 24 might be “summarized” by matching their trajectories to expected or ideal motions such as line segments, ellipses, “figure 8s”, or other planar cyclic curves—all of which experience shows generally appear in data sets from successful fitting sessions.
When relating the locations of markers 24 to each other (such as computing the distance between two distinct markers) at a given time, the time difference must be taken into account—especially for a marker tracker 28 which senses the location of only one marker at a time—even if the timestamps for the marker locations are only a few milliseconds apart. For instance, we might estimate the locations of two markers at exactly the time (which might be the average of the time stamps for those two locations, say). Such an estimate of the location of any marker at any given instant in time T must use non-linear interpolation, because linear interpolation for the cyclical movements would not yield the most accurate estimate. One well-known non-linear interpolation method is cubic interpolation. For it, take the marker's four coordinates with timestamps closest to T; compute the coordinates of the formulas for the 3-d cubic polynomial curve (parameterized by T), which passes through the four points; and then determine the point on that curve which corresponds to time T. This is a well-known technique in such disciplines as computer graphics.
Data analysis may include but is not limited to the following statistics:
- vertical hip motion range: The marker mounted to the hip is tracked for up and down range of motion, where ranges which are too small or too large suggest that the saddle height is not properly adjusted. It can also indicate that a different saddle altogether is appropriate.
- wrist motion range: The marker mounted to the wrist is tracked for any motions that stray significantly from a single point. If the wrist is moving during cycling, it indicates that the arms of the cyclist are over-extended or under-extended. Either of these conditions suggests that the handlebars are not in an optimum position for the cyclist, and should be adjusted or replaced.
- lateral knee motion range: The marker on the knee should describe a repetitive nearly planar pattern. Deviations from a plane indicate that the knee is swinging outward or inward from a “same-plane” position and that adjustment is needed (such as wedges between the shoe and the pedal).
- ankle motion range: Similarly, unexpected inward or outward movement (toward or away from the frame) with respect to the foot indicates that wedges or spacers are advisable.
- foot motion range: The foot marker should describe a nearly perfect circle; the foot should be pedaling pedals that themselves are traveling in a circular motion. Any deviation from nearly a planar circle suggests that foot wedges or other adjustments are needed.
- ankle-knee-hip angle range: Mean and extrema values for the knee angle range which are too large indicate that the saddle is too high; values which are too small indicate that the saddle is too low.
- mean knee-hip-shoulder angle: The mean value of this angle relates to the goal of the cyclist: small angle for speed, large for comfort; the invention might suggest certain angular ranges corresponding to various goals based on historically accumulated data.
- relative forward-aft location of knee and pedal: Non-zero forward-aft horizontal displacement of the knee marker with respect to the foot marker indicates that the saddle is too far forward or rearward.
Recommended ranges for the above should not be applied to an unconventional bicycle, such as a recumbent bike, of course. A separate set of optimal geometry statistics and recommended ranges must be kept for this and for other classes of unconventional bicycles. Even for conventional bicycles, separate optimal geometry statistics should be kept for different classes (goals) of cyclist: competitive road racers, competitive mountain racers, “serious” road or mountain bike enthusiasts, recreational cyclists, commuters, subclasses of these, and so forth.
Once a bicycle fitting session has been completed, and all variations in motion of the cyclist on the bicycle are found to be within the proper limits, the data from that session is added to a master database of successful fit sessions. For privacy reasons, the names (addresses, phone numbers, . . . ) of the cyclists would not be recorded therein. Hence, the more fitting sessions that are performed, the more honed the recommendation process becomes. That is, rather than using fixed, predetermined rules for making recommendations, the system can “learn” as it is updated with additions and modifications to the master database. Data from multiple systems can be pooled, analyzed, and shared to more quickly build quantitative ranges describing good versus bad fits for individuals of various body geometries. Session data from recognized expert fitters might be weighted more heavily.
Because cyclists differ substantially in size, some thresholds, limits, and recommendations should be scaled relative to distances between markers, which will presumably be proportional to the size of (the relevant body parts of) the subject cyclist. For example, the upper limit for inward-outward motion of the knee should be proportional to the leg size, which will roughly be proportional to the ankle-to-knee distance plus ankle-to-hip distance.
Once all data is gathered, it is analyzed for potential adjustments to bicycle components, substitution of different bicycle components, or even substitution of a different bicycle or bicycle frame. Without re-harnessing the cyclist, one or more acquisition sessions are repeated after the adjustments until the desired quality fit is achieved. Each of the measured statistical values can be re-measured and evaluated individually to optimize each aspect of the bicycle fit.
The detailed procedural steps for the preferred embodiment of the present invention are listed in FIGS. 2 and 3. FIG. 2 lists the steps that the operator of the system generally would perform manually (although the software 34 might at least prompt the operator for each step to perform and prompt the operator to input session and cyclist parameters). FIG. 3 lists the automated steps that would normally be implemented in or controlled by the bicycle fitting application software 34 running on computer 30. The division of tasks between the operator (FIG. 2) and the invention hardware and software (FIG. 3) is not absolute; an advanced system might do more steps automatically.
In FIG. 2, the method begins with step 101, which attaches detectable markers to at least one side of the cyclist (rider, person) and optionally to the bicycle (or other vehicle or conveyance being powered by the person). If additional markers are also attached to the bicycle, those markers can serve to define a coordinate system relative to the bicycle rather than relative to the tracker which detects and measures the locations of the markers. In either case, step 102 defines the axes of a three-dimensional coordinate system.
Step 103, which checks the orientation of the saddle or seat may be either manual step or may use a probe tracked by the tracking system to automate partially the check. This might also involve noting the style and width of the saddle. This check would be ignored or would use a very different criterion for a recumbent bicycle.
Step 104 acquires details and characteristics about the rider (cyclist) and the bicycle (or other conveyance). This would presumably include at least the name of the rider so that saved information can later be retrieved by name. Other pertinent information might include the cyclist's gender, age, experience (from beginner to racer) and goal (comfort, endurance, or performance efficiency). Other relevant information would characterize the bicycle. Because a 3-dimensional tracking system can also track markers on a probe, it is possible to arrange for such a probe to measure certain landmark locations on the bicycle such as the axis of the pedals, the top of the seat, wheel centers, points along the top tube or handlebars, and the like. More preferably the computer system would have a database of bicycle models and their geometry, so that once the bicycle model is identified, all its geometry would be available.
After directing the cyclist to begin powering the bicycle (step 105), which would normally be well supported in a stationary but otherwise normal mode of operation, step 106 would invoke the computer system to begin executing steps 201 through 208 in FIG. 3. (Note that the computer system may already be prompting the system's human operator to perform the manual steps of the method: namely, those listed in FIG. 2.) Step 201 sets up a marker tracking system to acquire a sequence of discrete locations at discrete times for each marker for a period of time. Step 202 then actually acquires and saves the sequence of locations for each marker. Along with the spatial coordinates are saved the time of acquisition, the success or failure of acquisition, and some association between markers and their corresponding temporal-spatial coordinates, such as an index number or mnemonic identifier. The acquisition time might be as short as a few seconds to as long as many minutes, but should include at least a number of sample locations over at least one cycle of the motion which powers the bicycle or other vehicle (such as at least one revolution of pedals or the complete cycle of a rowing stroke).
Optional step 203 transforms the coordinates to a more computationally convenient or standard coordinate space, such as a coordinate system with the origin at the center of the circle approximately traced by the foot marker and one with, for example, the +Z coordinate axis pointing up and the +X axis pointing forward. Similarly, the time coordinates may be adjusted to be relative to the beginning of the sequence of time coordinates.
In order to summarize the location information about each marker, step 204 computes statistics about each marker. This likely would include at least its mean XYZ location (centroid) and range of motion along each of the X, Y, and Z axes (corresponding perhaps to forward-aft motion, lateral motion, and vertical motion).
In a similar vein, step 205 would compute statistics about the distances between certain specified pairs of marker locations (each at the same instant in time). For example, one typically significant statistic would be the mean distance between the hip marker and the wrist marker. The most straightforward way to compute that would simply be the distance between the hip mean location and the wrist mean location. Another significant statistic would be the horizontal forward-aft displacement of the knee marker relative to the foot marker (such as the difference between their X coordinates). Changes in the pedaling cadence would also indirectly indicate the relative comfort and effort.
Still in the same vein, step 205 would compute at least one angle formed by a specified trio of markers. One obvious example is the angle of the bent leg at the knee and how it relates to proper saddle height. If the maximum angle is too near 180 degrees the seat is too high; if it is too small, the seat is too low. The leg angle is approximated by the angle formed by the ankle-knee-hip trio of markers at some moment. The mean value of this angle is probably not as interesting as the maximum angle.
Once sufficient discrete locations are acquired for each marker, each marker's sequence of locations may be fit to a continuous trajectory which approximates the actual path of the marker. This may be as simple as a sequence of connected straight line segments or may be represented by smooth curves such as cubic-polynomial splines. Although not absolutely necessary, there are two reasons for fitting the discrete locations to a continuous trajectory: (a) for a displaying visually pleasing graphical presentation of the marker paths to the system operator and (b) to provide the capability to interpolate locations of a marker at times between the times of two or more known, nearby acquired locations. The latter is particularly important if the location of only one marker is determined at a time. Interpolation allows the system to estimate the position of some or all of the markers at the same instant in time even though the locations of the markers were acquired at slightly different times.
Furthermore, the actual trajectory might be matched with or compared to an ideal or standardized trajectory or to all historically recorded marker trajectories of the same cyclist.
Step 206 of FIG. 3 allows for the optional visual depiction at least some of the acquired and computed data overlaid onto a stylized graphical representation of the person (cyclist) and vehicle (bicycle). Alternatively, as shown in FIG. 4, the data may be in a tabular form with reference identifiers associating each datum with the distance or angle identified on the graphical representation of the cyclist. Step 207 further provides for a permanent printed or electronic report that the can be kept for future reference or given to the person. This may include the original acquired coordinates or simply the computed results and recommend changes or both. Similarly, the computed results (angles, distances, their ratios, and like statistical results) can be compared with norms or with results from previous sessions (step 208).
The computer software would compare the results of the current session with a database of norms, measurement ranges, and other criteria, where the database is built from experts' advice and from collected results of successful fittings. The software would match the present session results with entries in the database. Associated with entries in the database would be recommended alterations to the bicycle (or other vehicle) or its components. Alternatively, for somebody desiring to purchase a bicycle, the database would list possible models, frame sizes, and changeable components which best match the present results. This could be implemented using common database software. Obtaining the data content for the database is the bigger practical challenge.
Because all fitting data is available on the computer system, over time it is available for compiling into statistical norms which would form a growing and improving content for the database. Data from many installations of the invention would be centrally compiled and shared. It would be important to attach to each cyclist's data file a rating indicating how satisfactory the cyclist found the fit to be. Ideally this assessment would be made after a period of real-world experience riding the bicycle.
Once the computer system has acquired motion data, processed it into meaningful summary statistics, and supplied database-oriented recommendations, the level of satisfaction is determined in step 109 of FIG. 2. If the statistics are within norms that match the cyclist's goals (comfort, endurance, efficiency or the like) and if the cyclist is satisfied with the quality of fit with the bicycle the session may be completed simply by saving acquired data and/or computed results in a form for possible future reference. The results would then be used to design a custom bike, finalize purchase of the bike used, finalize purchase of the new components used, or aid in some other decision for which the invention was used to quantify the quality of fit.
If data was acquired from only one side of the cyclist, it might be desirable to acquire data from the other side and compare the results for equality or symmetry, for example. This could be done whether or not the results suggested that the cyclist-to-bicycle interface (fit) was satisfactory. The latter suggests that the computer software be capable of overlaying the results for both sides for visual comparison—this in addition to the obvious item-for-item comparison of the numeric results. See step 108 of FIG. 2.
However, if the last report of results was unsatisfactory or the computer software recommended a substitution or a bicycle-specific alteration to the configuration, the cyclist or operator can choose whether to comply. See step 107 of FIG. 2. After making the change, the whole method would be reapplied—presumably returning to step 105 (if not an earlier step). The alteration would be tailored to the cyclist's goal (comfort, endurance, maximal power output, or so on) and skill level (beginner to professional). Potential alterations (besides substitution of a wholly different frame or bicycle) could include any of the following: longer or shorter pedal crank arms, pedal spacers or wedges, seat width, seat height, seat angle, seat fore/aft position, handlebar height, handlebar stem length, handlebar style, and handlebar angle. If the cyclist is riding a bicycle simulator instead of a normal bicycle, frame geometry and dimensions can also be adjusted (manually or perhaps even automatically under computer control)—as if a new bicycle was being substituted.