BIOMECHANICS OF BICYCLING

Transcription

1 BIOMECHANICS OF BICYCLING Rory R. Davis Structural Dynamics Research Corporation San Diego, California M. C. Hull Department of Mechanical Engineering University of California Davis, California ABSTRACT A computer-base ins trumentat i on sys tern was use to accura te1y measure the six foot-peal loa components an the absolute peal position uring bicycling. The instrumentation system is the first of its kin an enables extensive an meaningful biomechanical analysis of bicycling. With test subjects "riing" on rollers which simulate actual bicycling, pealling ata were recore to explore four separate hypotheses. Experiments yiele the following major conclusions: 1) Using cleate shoes retars fatigue of the quariceps muscle group. By allowing more flexor muscle utilization uring the backstroke, cleate shoes istribute the workloa an alleviate the peak loa eman on the quariceps group; 2) overall pealling efficiency increases with power 1eve 1; 3) non-mot i ve loa components wh i ch app ly averse moments on the knee joint are of significant magnitue; 4) analysis of pealling is an invaluable training ai. One test subject reuce his leg exertion at the peal by 24 percent. NOMENCLATURE Fx,Fy,F Z MX,My.M z e l e 2 x,y,z n (e l ) P.I. Force components acting on the bicycle peal (N) Moment components acting on the bicycle peal (Nm) Angle of crank arm relative to vertical (ra) Relative angle between normal to peal platform an crank arm (ra) Peal platform coorinate axes Instantaneous motive efficiency Pealling performance inex F max Maximum resultant of F x an F z (N) M ab Resultant magnitue of Ma' M b (Nm) M xyz Total resultant moment magnitue (Nm) ill

2 114 ± \ Angle of rotation about x axis (ra) ± 6 Z Angle of rotation about z axis (ra) ± X Translation along x axis (mm) T(e l ) Instantaneous torque (Nm) P Average single peal power level (W) L Length of the crank arm (mm) ca I NTRODUCTI ON A new bicycl ing measurement system escribe by the authors in Part I has enable extensive pealling mechanics aalysis. The new instrumentation, augmente by the capabi 1ities of special ize computer software for post-processing, was use to investigate several aspects of pealling. This paper presents the results of those experiments. First, several foot-peal connection types are investigate to etermine their effect on rier fatigue, pealling power, an efficiency. Secon, complete loa an peal angle results are presente an their relation to rier fatigue an overuse injury is iscusse. In aition, an attempt is mae to evelop a set of features that may be use to etect foot orientation changes on the ajustable peal/ynamometer. This exercise is the first step in the creation of a pattern recognition scheme for etection an prevention of bicycling overuse injuries. Finally, an experiment showing the value of pealling technique ana lys is in improv i ng pea 11 i ng performance is presente, along with other interesting finings. Instrumentation an Apparatus The measurement system consists of four sub-systems: 1) the six component peal/ynamometer, 2) the peal position measurement system, 3) signal conitioning electronics, an 4) the minicomputer which igitizes, stores, an analyzes transucer ata. The ynamometer, peal position measurement harware, an signal conitioning amplifiers were specially esigne for this application. The strain gage peal/ynamometer incorporating four octagonal rings is shown in Figure 1. All six loa component s on the pea 1 are compute from a set of eight inepenent Wheat stone brige circuits. Figure 1 also illustrates the potentiometer which measures the relative angle between the pea 1 an the crank arm. A s imi 1ar potent iometer is geare to the crank to measure the absolute crank angle. The loas an angles measure are efine in Figure 2. Transucers are esigne to fit any commercially available bicycle without moification so that subjects may rie their own bicycles in experiments. The entire system, incluing the signal conitioning an igital conversion electronics, is calibrate using a newly evelope software routine an a calibration table for precise application of loas to the ynamometer. The potentiometer signals are calibrate automatically uring each riing experiment. The calibration techniques are escribe more fully in Part I.

3 Data Collection an Analysis After the measurement system is calibrate, laboratory experiments are perfol'me with the test subjects riing on "rollers" (see Figure 3). Both wheels revolve an no lateral support is provie, so that riers must balance their bicycles just as if they are riing normally. Before riing begins, the ata acquisition program measures all static offset voltages of the system. When a test commences the rier peals against the significant roller friction until the computer operator is satisfie with the spee an steainess. The computer software operates interactively to continuously write the pealling spee to the operator. The operator triggers the acquisition program which begins acquiring ata at the beginning of the next stroke (peal top ea center). For most tests, the cycl ing spee is 8.37 ra/s (80 rpm) an ata is taken over three complete revolutions. Data channels are sequentially sample at 15 khz in bursts separate by a 1.4 ms interval. Due to sequential sampling, maximum phase error is 2.3xlO-3 ra (0.16 egrees) which is less than the resolution of the igitize potentiometer signals. The resulting single channel sampling rate is about 500 samples per revolution. The ata is store on har isk for later processing. Data analysis is facilitate by special software which performs the following functions automatically: i) Conitions an calibrates 91 an 92 ata. ii) Corrects for static offsets an iscretizes ata further to integral egrees of arc. iii) Computes the point-by-point average time history of output over three revolutions. iv) Calculates all applie loas using the previously obtaine sensitivity matrix. v) Calculates instantaneous torque an power applie to the crank. vi) Calculates motive efficiency as the instantaneous percentage of the total loa vector contributing to positive work (the component of loa in the irection perpenicular to the crank). vii) Calculates average power an performance inex. viii) Digitally filters the results to eliminate noise. ix) Presents results in screen plot form. In calculating torque, motive efficiency, an average power only loas Fx an Fz are consiere. The torque T an motive efficiency n at a point 91 egrees into the peal arc are efine by 115 (1 ) T (9 1 )/L ) n(9 1) = [2 2]/2 x 100 percent (2 ) ( Fx + Fz

4 116 Fig. I. Six loa componenl peal/ynamometcr.,.%j Fig. 2. Definition of ynamomeler loa components an polenliomeler angles.. Fig. 3. Riing on rollers in the laboratory. Fig. 4. Three ifferent root-txal connections.

5 where Fx an Fz are functions of 91. Note that, as efine, negative efficiencies may result when torque on the crank is negative, against the irection of rotation. Motive efficiency inicates how well a rier utilizes muscle exertion to power the bicycle, while torque presents the magnitue of useful work. Average power P in Watts, ID (3) is also a magnitue measure, convenient for scaling the loa histories of one run to another if esire. It must be note that average power is given only for one leg, an is measure at the point of application of the loas. Since most power measurements were mae at the rear wheel in other stuies, one must be cautious when comparing results. A quantitative measure of overall efficiency of converting peal loas to useful work is given by the performance inex. Optimal pealling force performance woul be attaine if a rier coul apply a maximum constant force perpenicul ar to the crank arm for a complete cycle. The performance inex scales the actual rier to that optimum by the equation 360 P.1. =( T(9 1) max (4) where Fmax is the maximum resultant of Fx an Fz exerte uring the cycle. Due to the 360 Fmax scaling, perfect performance yiels a P.I. equal to unity. The P.I. is particularly useful because the normalization allows irect comparison between riers without the nee to consier absolute power output. RESULTS A. Foot-Peal Connection Stuy Bicycle riers have long realize the relationship between the foot-peal connection an fatigue an a three-stage foot-peal connection hierarchy has evolve. In the first stage, force is transferre to a bare peal through a soft-sole shoe (see Figure 4(a)). Fatigue occurs reaily because the only muscle groups active are those which exten the leg. The work of Houtz an Fisher (1959) supports this argument. The next stage as. toecl ips to the peal. Toecl ips constrain the longituinal an transverse position of the shoe as shown in Figure 4(b). Also, a leather strap applies a pre-pressure between the shoe sole an peal. The toecl ips allow more muscle groups to participate in moving the peals. In the electromyogram stuies of Tate an Shierman (1977), ata recore with toeclips showe that flexor muscles were active uring the upstroke an the extensor muscles were active over greater arcs. The th ir stage retains the toec1ip but replaces the soft-so1e shoe with a special light-weight shoe reinforce with spring steel in the sole (see Figure 4(c)). A cleat fits the peal so that both longituinal motion an ax i a1 rotat ion of the foot are prevente. When the toec1ip strap is pulle taut the foot cannot be isengage. From a muscle utilization

6 118 stanpoint, this situation offers clear avantages because more muscles can supply motive power over the peal arc. Since electromyogram ata for the thir connection are not in the literature, experiments were performe to ocument all three connection types an to verify previous finings. Foot-peal connection experiments were performe early in the testing program. Data was sample sequentially at 1000 Hz per channel an only a single revolution was acquir-e uring each run. Several runs uner similar conitions verifie that major features of the results were repeatable, however. All testing was performe at a pealling spee of 8.0 ra/s (76 rpm) with a single peal power level of about 125 W. The rier was an experience long-istance bicyclist an was not tol to peal in any special manner. Torque an motive efficiency plots for each of the three foot-peal connections are shown in Figure 5. Inspection of the results in Figure 5 gives insight into the pealling process. Several major observations can be mae: i) Use of the cleate shoe greatly improves pealling efficiency. Note the numerous peaks in efficiency as oppose to the soft shoe results. The aitional peak from 190 to 240 egrees inicates an active role of the cleat in power generation. Inspection of loa ata reveals that larger negative shear loas are prouce uring the backstroke an the peal is positione to utilize that loa more effectively. Better plantarflexion control an work by the hamstring muscle group are mae possible by the cleat. ii) In both toeclip cases, aitional efficiency is achieve early in the pealling cycle uring the first 25 egrees of arc. This is primarily ue to greater orsiflexion which enhances normal loa utilization. Aitionally, the toeclip provies greater shear (Fx) capability. iii) The maximum torque level ecreases with the aition of toeclips an cleats. For the shoe sans toeclip, negative torques exist over 160 egrees of peal arc an the energy lost must be regaine by increasing normal peal loa (an torque) in the first half of the cycle. When the toeclips are ae, improve efficiency early in the cycle lessens the epenence on extensor muscles to provie the major motive work, esp ite the 1es s severe, but still present negat i ve torques occurri ng in the backstroke. iv) Torque history changes ramatically with the aition of cleats. Maximum torque is lower since efficiency in the backstroke couples with significant negative shear loa to prouce positive torque past 200 egrees of arc. In fact, negative torque occurs only for 55 egrees of arc with the cleate shoe. Clearly, the cleats allow enhance activity of the flexor muscle groups. Initial testing showe that fewer ata points per revolution were aequate to accurately escribe the loas. Therefore, the sampling algorithm was altere to sample at a slower rate while maintaining small phase error with the "burst" sampling technique escribe earlier. The remainer of the tests were performe with the right peal instrumente an the slower sampling enable acquisition of 3 complete cycles which were point-by-point average uring post-processing. Toeclips an cleats were use by the riers an all ata were taken at 8.37 ra/s (80 rpm).

7 119 'j ",0 D,,0'j, ",',, ci ON CC o,; ;.,--;5;-:0'"".-:-:10::0'"".--:-l"50'"",--=2""0=-0.---'2'"'5"'0-.---::"30"'0-,-3-50'-, CRAN ANGLE. DEGREES Ne > u ",; f ::> u ;:ci :t,t o N 7 0,+.-""5;-:0'"",-'-10'"'0'"",.1"!"50: , CRANK ANGLE. DEGREES (a) soft sole shoe ",0 '"', ci ON CC o.,,0 0" " w',,,;r ""<:::"""---==- _-_--_---_-"!"c:_ '0. SO. loo. lso. 200, CRAN ANGLE. DEGREES,. cl u z,; t '>,;: u ci T,; 7 0 ".-.".,,.--.,!"":""-=---=, ,,- so. loo, ISO , 350, CRAN ANGLE. DEGREES (b) soft sole shoe with toeclip '"', wo,,,0 0" " w',, w, =>0 ON '" o,; f =>-=".-"""===---,; f ::::."<:::"----- Ol-.-=--,,::---:-:o-::---:::-:---==---,=-== SO. loo, S CRAN ANGLE. DEGREES,; N ><,; ':" u z, u c:i t j- 'fo.-"'50.-1""0-0.-1'!"so , CRANK ANGLE. DEGREES (c) lightweight shoe with cleat an toeclip Figure 5. Torque an motive efficiency for the three foot-peal connections at a single peal power level of 125 Wan 8.0 ra/s (]6 rpm)

8 120 A number of riers performe in the tests that follow. be escribe as: Briefly, they can Rier A: Rier B: Rier c: Rier D: Experience amateur racer, places well in long-istance races, Male. Alternate for U.S. Junior National Team, Male. One of the top female racers in the U.S. Inexperience bicyclist, Male. Rier E: Experience long-istance bicyclist (was subject for foot-peal connection stuy), Male. Rier F: Experience long-istance bicyclist, Male. B. General Peal Loa Results Figure 6 presents the complete loa profi les for rier A with a single peal power level of 180 W. The general features of a typical loa profile are containe in Figure 6 an will be iscusse. Figure 7 contains the peal angle, torque, an efficiency plots for the same run. The most prominent feature of the loa profiles is the shape of the normal loa (Fz) curve. As in Figure 6(a), the loa reaches a max imum magn itue in the range of 90 to 110 egrees an then the magnitue ecreases towar zero. The two maxima uring the backstroke ( egrees) almost always occur an excursions above zero of significant magnitue are rare. The ip between maxima always occurs near 270 egrees, where the horizontal velocity of the foot changes irection. Therefore, the ip may be ue to inertial loas. Another notable feature of the loaing is that Fy an Mz are always in phase with the normal loa Fz. They all reach maximum magnitues at about the same point. As the rier pushes ownwar in the maximum loa zone of the stroke, the foot also pushes outwar on the peal an the heel of the foot tens to twist outwar. The twist is a simple result of the moment cause by noncoincience of the point of application of Fy an the central axis of the lower leg. The origin of Fy is not as easily ascertaine but is likely to be cause by the geometry of the human boy relative to the peals. Since the magnitues of Fy an Mz are significant (Fy is usually of the same orer as the ri vi ng force Fx), an since both exert averse moments on the knee joint, measures to ecrease or eliminate these loas may be beneficial. It is of utmost importance from an injury stanpoint that Fy an Mz are not negligible, as usually assume in bicycling analyses. The profile of the force Fx in Figure 6(c) is a typical one when cleats with toeclips are use. Positive shear (forwar) occurs uring the first part of the stroke. As the peal moves ownwar, the shear loa shifts to a negative value to maintain a positive torque contribution. It is interesting to note that the negat i ve of the Fx prof i 1e near ly matches the My profi 1e. It may seem surprising that My is not always zero because the peal platform is fixe to the spinle with low friction ball bearings. The evelopment of such a moment is mae possible by the interaction of the shoe, peal, an toeclip. The My moment is a necessary byprouct of the shear force Fx because the total moment about the peal spinle must be nearly zero.

9 .l. If,j s "'. H.: '" 51. IIlG lIA. 2SG. JOa. 35ll. w \! 51. lilg. l5l 2lIA. 2SG. 3OlI. 35ll. 6OD. CNlIIK IIIGlf. Dflllll!ES CIIAIIK IIIIU. Dflllll!fS (a) () a,. to " H ' I If ạ T ṣ :Rc.. :1 "', 51. IIlG lIA. 25a.,00.,sa. 'IlG. Ta. sa. loo. ISIl. 2llD CIl. nil. 'IlG. CftJlIIK IIIDlE. DflillHS CIUllIK IIIGlE. DfWES (b) (e),,; " ". ",j 11 ;; aḍ If z. (J.: " '". l[" a x' l[ "-c 1 ;::; 51. loo. ISa. 2lIQ. 25a. 3OG. 'Sa. 'IlG. 51. loo. lsll. 2lIQ. 25a. 3OG. na. 6OD. CftJlIIK IIIGlf. DEGlIHS CMIIK IIIGlf. DEWES (cl (f) Fig. 6. Typical peal loaing of rier A at180w single peal power level an 80 rpm.

10 122 Accoringly, resolving Fx an My about the peal spinle axis yiels almost zero moment at that point. Any small non-zero remainer is ue to friction in the bearings an rotary inertial loa of the peal itself. The most unpreictable loa component is the moment Mx. Mx is generate whenever the point of appl ication of Fz moves 1aterally from the center of the peal. Such motion epens on several highly variable factors, such as 1ateral constraint of the foot an conition of the shoe cleat grooves. Therefore, large variations of Mx profiles are observe among riers. From an injury perspective, Mx is relatively unimportant. For example, the moment cause by Fy on the knee in the same irection as Mx is a maximum of 36 Nm for rier A, seven times the maximum value of Mx. Several other etails of the interaction between the ifferent outputs merit iscussion. Most importantly, note that efficiency reaches 100 percent before the normal loa Fz reaches its maximum magnitue (see Figures 6 an 7). At a crank angle of 75 egrees, the peal platform is nearly horizontal, not parallel to the crank arm. Due to the positive shear loa, however, the resultant loa is perpenicular to the crank an, hence, 100 percent efficiency is achieve. Since the maximum resultant loa occurs some time after 75 egrees when efficiency is sub-optimal, rier A has waste some of the maximum loa available. A secon point of interest occurs at a crank angle of approximately 220 egrees where the efficiency curve exhibits a secon peak. High efficiency occurs because the peal platform is nearly perpenicular to the crank arm an the shear force Fx is twice the normal force Fz. During the upstroke, the majority of torque is prouce by i nc 1in i ng the peal an pu 11 i ng back rather than pulling vertically up on the peal through the leather strap. One feature of Figure 7, which is not consistent among all riers, is the high efficiency profile. In comparing Figure 7 to Figure 5, also note that rier A manage to maintain positive torque for the entire cycle. It has been foun that such performance is achieve by subtle changes in peal angle an loa profiles. The preceing iscussion inicates the importance of proper peal orientation in achieving high pealling efficiency. Importance of peal orientation is well known by competitive cyclists who employ a process calle ankling to control peal position. Mastery of the technique, however, is not a simple process since riers normally have no irect metho of performance evaluation available. As a result, overall peall ing efficiency varies wiely among riers. Table 1 presents the best performance inex for each of the riers uring the main testing sequence. Note that riers E an F have similar riing experience an roe the same bicycle uring tests at similar power levels, but the P.!. levels are ifferent by about 15 percent. Furthermore, rier A's P.!. is 15 percent to 20 percent greater than rier E's, although the same bicycle was again use. Riers Ban C, using their own bicycles, yiele low P.!. values espite their extensive riing experience. This contraiction may be a result of the low power level for their tests relative to the others.

11 123, 5G. lm. lsg. Zllll. 25G. 3CIO. S5G. a. rniiil IlIEU. DfllllH5 11 :: c.; I. It. g L. ". _._ lm ZlIll. 25G no. UIl. C1llIIIK IlIGlf. DfllllHS Fig. 7. Companion torque, motive efficiency an peal angle for rier A.

12 124 TABLE 1. Best performance inices achieve by test subjects at 8.37 ra/s (80 rpm) Rier Power Leve1*, W Performance Inex A+ B C D E F * Single peal + This was achieve uring the run shown in Figures 6 an 7. Further evience suggests a relationship between power level an performance inex. Examining the test results of the inexperience rier D shows that the P. I. is greater than that of three of the more experi ence riers. The high power level may be a factor in this case since rier D maintaine a power level about twice as great s riers Ban C. Power level ifferences were ue to bicycle tire pressure an type, lubrication, an rive ratio. C. Power Level Effects on Efficiency Rier D was chosen for these tests because a more experience rier woul probably utilize special ank1ing techniques at the higher loas. The objective was to observe whether efficiency improvement occurs naturally with greater power level. The results of tests at three ifferent power levels are superimpose in Figures 8 an 9. Power level was varie using the bicycle's gears, while holing cranking spee constant. Note that as power level increases, efficiency in the 180 to 300 egree range improves significantly. The performance inex graually increases from for the 85 Wtest to an for the 140 an 200 W tests, respectively. Therefore, increasing power output by a factor of 2.4 requires only a factor of 1.8 increase in leg exertion at the peal. 1nl i ght of the fi n i ng above, it may appear that a ri er cou 1 exten bicycling range by riing at a higher power level. However, the resistances to motion uring actual bicyc1 ing must be consiere. Air rag force, the major eterrent in bicycling, varies with the square of the spee, an hence power varies with the cube (see Whitt an Wi1son (1974)). Therefore, the

13 125 cl :i ", :! 1< Z.,. :'i >-9 "-, :i "!: 5i ". :i. )( " W ow--, - 200W 158. ZII. :l5e. _ Aaf. Dfftn (a) 15G. 2tI c.- IIICl.f. DflllffS (b) oi E.. ". e". 1< ".; 58. IIlll :l5e. liilil.f. Dfftn () j.; i. :i,. =ci.. '",; :; " " ' 'f. o ". ẓ z. a,: za. za a z ci, :,',, -, I a 5G. IIlll. 15G. 21lG, lI. 4Illl. tllllllll AI&l.f. DE&llffS (f) FII- 8. LOl profiles 0( rier 0 I1 Ihrcc ill.reol po"r kvel.

14 126. M. lj -- 85w ow --200W Io' ;1.' ḍ. 10: so. ldo iIG. 2SG.!Oa GO. 2. >:,.j :I :" Ḍ. so. IDO D! CIIIINK AMGtf. "..,0 c 0 D,.j -' O' Zo et" -' et. co L loo. Iso Q. 3Oll. 3so CIIIIIK AMGtE. OEGftfES Fic.9. Companion ' que. IDOli.ffincy an pealoncl< for rier 0 ot three illerenl power 1e1.

15 factor of 2.4 increase in power woul result in a factor of 1.34 increase in roa spee. Since the force exertion factor is greater than the spee factor, total range is ecrease assuming a fixe amount of energy is available to the rier. The situation is analogous to an automobile, where fuel consumption increases isproportionately with spee; although in the bicyc1 ist 's case, the power p1 ant efficiency (performance inex) increases with power level. The cause of the increase in P.I. for higher power levels may be euce by inspecting the profiles. Peal angle changes little uring the backstroke but the magnitues of Fx an Fz are quite ifferent for each test. Negative shear in the backstroke increases significantly an the backstroke maxima of the normal force become more pronounce. Note the lobe in Fz evelope uri ng the 200 W tes t near 230 egrees. Although the lobe never excees zero, the benefits of less negative torque ue to normal loa are reflecte by higher efficiencies. The torque plot shows negative torque occuring over about 100 egrees of arc instea of 140 egrees as in the 85 Wtest. Figure 8(f) shows that the Mx moment is inee quite ranom, even between tests with the same rier. Accoringly, the moment appears to be primari ly ue to inaequate lateral foot constraint by the leather strap. The Fy an Mz plots also show some interesting behavior. Although both loas track the normal loa Fz in phase, the magnitues are not relate to Fz in any simple manner. Fi na lly, the si mi 1arity between Fz an the torque shows that the majority of power increase is provie by greater normal loa (extensor activity) uring the ownstroke. D. Peal Orientation Experiments Extensive testing was performe with riers A, B, e, E an F to investigate the effect of foot orientation on loa history. Each rier performe a minimum of 7 tests which inclue positive an negative ajustments in three irections an a normal ajustment test for comparison. Specifically, ajustments inclue :I: 3 egrees rotation in the z-irection (:I: ez), :I: 5 egrees rotation in the x-irection (:I: ex), an :I: 0.5 cm translation in the x-irection relative to the peal spinle (:I: X). Figure 10 presents the loa profiles of rier A for the normal test with results for positive an negative ajustment of e z superimpose. Although changes in the loa profiles of Figure 10 are not large, there o appear to be some features of the loaing that istinguish one ajustment from the other. For example, the value of Mz increases or ecreases uring the secon half of the stroke epening on whether the ajustment is negative or pos it i ve, respect ive 1y. Other features concerne wi th peak or range magnitues may be selecte which yiel meaningful istinctions between ajustments. It shoul be note that Figure 10 is presente as an example but is not typical of tests with other riers. It was foun that, in most cases, visibly usefu 1 features extracte from one rier' s resu 1ts were contra icte by the results of some or all of the remaining rier tests. Aitional computer software was therefore evelope in an effort to ientify representative features not iscernible by visual inspection. 127

16 128 J J :11' 0" le'.,.:. J \' N "-; J cl "I I 5lI UlL l5li ZlIll. 25lI 3OG wo. CllAIIK AI6l.E. DEWES (a) a.ft :11 D J fa.\'.: J 1 >lj "-,. " l I \: SD. loa. l5ll ZlIll. 25lI 3DG.!So. CRlItIK All6l.E. DEIlftfES (b) g J u C :a DJ :a a D, ;j. 1 I I,. --- j- 'I I - J "I \: 50. lllll. lsd. 2UD loa.!so. CIIAIlK AllGlE. DE61lEES (c) " l<. z J N :c " c : :11. 0" "..: 1 D " >!l :c oi. J z 1 si lllll.l5li2ll1l.25li3og.s5liwo. CIII1IK AllGl.E. DEIllIEES () z, t: S ldil. lso. 2UD DG.!So. C!VIIIK AllGlE. DEGREES (e) M ",. ẓ. ' r: o '",.,,. SD. ldil UD DG. 15l1 C!VIIIK aigl.e. DEGREES (f) Fi&- 10. LOB profiles of rier " lor Ihr«ial fao' rolalion aju'lmcnls

17 The three moment components were affecte most by foot orientation ajustments, so subsequent feature calculations were performe on Mx, My an Mz only. First, average an RMS (root mean square) values of the moments were calculate for three intervals of peal arc: zero to 180 egrees, 180 to 360 egrees, an full cycle. In aition, all Rermutations of resultant magnitues of the moments (e.g. (Mx2 + My2)1/2 = Mxy, etc.) were calculate as functions of peal arc. Finally, the resultants were processe to yiel their average an RMS values over the three intervals iscusse. Unfortunately, the calculations faile to ientify any loaing features representative of all riers for a particular foot-peal ajustment. However, some features were foun to be common among groups of two or three riers who were teste at similar power levels. For example, the RMS value of the tota1 moment resu ltant (Mxyz) over the secon ha 1f of pea 1 arc was foun to be a iagnostic feature for 6z ajustment for riers A, E, an F. These three riers were teste at similar power levels as Table 1 inicates. While several features were ientifie as iagnostic for subsets of riers, assembly of enough inepenent features to separately ientify the ajustments was not possible in this investigation. E. Performance Briefing Experiment As iscusse earlier, results of bicycling tests showe that pealling performance inex varie greatly among the riers an the P.I. i not appear to hol any corresponence with cyclist experience. Both riers an investigators hypothesize, however, that pealling efficiency an hence P.I. coul be improve, especially when the appropriate feeback was returne to the rier. Therefore a simple experiment was performe to investigate the effect that iscussion of pealling technique has on rier performance. In initial phases of testing before any iscussions, rier E ha performe several runs at a power level of 160 W. One typical run was use as a control case an rier E became the subject of the experiment. Before subsequent testing, rier E was briefe concerning his pealling technique with the use of the soli 1i ne control case ata in Fi gure 11 (enote "before"). The most prominent flaw in the cycle was the occurrence of negative torque uring the backstroke. To remey the situation, rier E was instructe to increase both normal pu 11 an negative shear on the peal uri ng the backstroke. Si nce the shear loa was 1arger than the normal loa in the original test, rier E was also avise to use greater "ankling" to better utilize the shear loa in the 200 to 270 egree range (calf muscle contraction). All these moifications woul help raise the efficiency curve (see Figure ll()) in the latter stages of the cycle. A final recommenation was mae to shift the main peak in efficiency to correspon with the peak in Fz magnitue. Although this moification woul not increase efficiency, it woul inuce larger terms in the numerator of the P.I. expression (Eq. (4)). Calculations showe that the efficiency shift coul be achieve by altering "ankling" behavior to exten the toe earlier in the stroke. The ashe line ata in Figure 11 shows the results of a test of rier E at the same power level after briefing. The improvement in backstroke efficiency is ramatic an the resulting torque output remains positive over the entire cycle. Whi le peal angle shows improve calf activity in the

18 a lil \: I sa. LcI :,., 1U. l5l Z54. 3Oll. m. _ a cl sa. IlL L54. 2lIG. Z54. 3Oll. JSG. Ull. CftAI1\ IlIIU. DfWf5 CIflIlK HliU. DfWfS (a) () l-- " -. z za a :I =clj_ zcl., / sa ". ci r, cll IV a se IIlG. L54. 2lIG. Z54. 3Oll. SSG. Ull. ;: sa. loa. L54. 2lIG. Z54. 3Oll. I5G. Ull. ClIIMlK HGU. DfWf5 CftAI1\ HIU. DfllllH5 (b) (cl 1 gl Ij - /1 tj, Eo r >c. "-0,. ". a1 E" c C a,.:").j za c' et" I. sa lii. l5l 25L JSlL '11 sa. llill. L54. 2lIG. Z54. 3Oll. SSG. Ull. CftAI1\ '"KU. DfllllH5 (c) ClIIIIK IIIIU. DfWf5 (f) Fig. 11. Peal force an other results before an after pealling technique analysis of rier E.

19 to 270 egree range, the vast majority of ae efficiency an torque is ue to altere loa profiles. Rier pull on the peal is quite pronounce in Figure ll(a), where Fz remains positive for over 40 percent of the total peal arc. Such a result is unpreceente in the literature concerning bicycl ing measurement. The much greater positive normal loa tens to al ign the total loa vector more perpenicular to the crank arm an hence increases efficiency. In aition, Fx shows larger magnitues in the 200 to 300 egree range, which translates into higher torque in the backstroke. A conven ient by-prouct of the improve backs troke performance is the ecrease of peak magnitues of all the force components. Constancy of the power level requires that total area uner the torque curve remain constant assuming no variation in angular velocity. The increase in torque from 200 to 350 egrees cause by enhance flexor muscle activity rel ieves the loa necessary from extensor groups in the 50 to 100 egree range of maximum exertion. The improvements gaine by briefing raise the P.I. value from 0.29 to 0.38, which translates to a savings in rier leg exertion at the peal of 24 percent. CONCLUSIONS The results presente here both confirm the conclusions of previous research an she new light on the pealling process. The results of the foot-peal connection stuy support the finings of Houtz an Fisher (1959). Also, Tate an Shierman's (1977) claim that more muscle groups come into play with toecl i ps is substant i ate. However, toecl i ps alone o not prouce an increase in the amount of peal arc that experiences pos iti ve work. The primary function of the toeclip is to increase efficiency by permitting greater orsiflexion an shear loas in the initial stages of the pealling cycle. Aitionally, it was foun that cleats allow a marke increase in flexor muscle activity, especially uring the backstroke. Contrary to popular belief, negative shear loa rather than positive normal loa is most important uring the backstroke. The normal loa is almost always contributing negatively to the torque uring this phase. By proucing the highest pealling efficiency an alleviating the extensor muscle loas, cleats appear to help ecrease lower extremity muscle fatigue in bicycling. General attributes of loa profiles were foun to be similar among riers (except Mx), an greater efficiency was achieve by subtle changes in the Fx, Fz an peal angle profi les that appeare to cost 1ittle in terms of rier energy expeniture. The magnitues of Fy an Mz were foun to be significant an measures to eliminate those averse loas will probably be beneficial in alleviating rier knee injury. Rier injury an fatigue are likely to be minimize by moifying pealling technique to more evenly istribute muscle requirements throughout the lower extremity. The significance of performance inex in relation to rier fatigue merits aitional iscussion. It shoul be note that in the strictest sense, P.I. measures the ri er' s overall abil ity to convert pea 1 loas to usefu 1 work uring pealling. Therefore, the P.!. oes not irectly involve the actual rier muscle exertion, only the en result of muscle exertion (loa at the peal). However, the following observations lea to the conclusion that P.l. can also inicate the actual muscle exertion of the rier: 131

20 132 1) P.!. can be increase greatly by a re1ati ve ly small vari ation in pealling technique. The muscle energy cost of varying the technique is therefore small, while the greater P.!. enables significantly enhance power output for comparable loa magnitues. In other wors, the main effect is the shifting of existing loas to more useful orientations. Therefore, neither muscle exertion nor loa magnitue are greatly affecte. 2) The peal loa components important for torque, efficiency, an P.I. calculations are closely relate to the action of the larger leg muscle groups (e.g. flexors an extensors). It woul therefore be reasonable to assume that the peal loa components are goo inicators of overall leg muscle exertion. 3) Some leg muscle work is not reflecte in the peal loa ata an subsequent calculations. For example, the energy cost of lifting the leg against gravity (when one tries to pull up on the peal in the backstroke to improve efficiency) is not consiere in the P.I. However, these types of exert ions are small in compari son to the main leg muscle action. Consequently, assuming that P.!. inicates muscle exertion is reasonable. Although the use of the P.!. for overall muscle exertion stuy is only approximate, it is felt to be of practical utility in pealling performance evaluation. An increase in P.!. at a specific power level will most likely give a ecrease in muscle exertion. Further stuy on the relationship between performance inex an rier exertion is warrante, however. Performance inex was foun to vary appreciably among riers an results show that pealling technique is a complex subject not easily unerstoo an mastere by even experience riers without avance ass i stance. I t appears that successful cycling is presently preicate upon evelopment of stamina an strength with a lesser emphasis on peallng technique an efficiency. Overall pealling force-to-torque conversion (P.I.) was foun to naturally increase with power level but air rag effects in bicycling preclue the ability to exten cycling range by increasing spee (an power level). Peal orientation testing an analysis faile to prouce the features necessary to accurately iagnose foot orientation ajustment. However, some features were foun to be iagnostic when power level was coltltlon among the tests. Therefore, aitional analytical an experimental research may unearth the features neee to complete the classifier. The problem is complex an warrants more intensive stuy. The specialize measurement an analysis system utilize prove to be an easily use, extremely valuable tool for bicycling mechanics analysis. Avance analysis capabilities enable the timely output of accurate an useful ata such as torque an efficiency profi les, an most importantly, performance inex which conenses a complex set of outputs into a single pea11 i ng techni que measure. The uti 1ity of such output ata was proven by the outcome of the performance briefing experiment, where a simple iscussion of technique resulte in a 24 percent savings in rier leg exertion at the peal. The possibilities for rier training to improve energy utilization an minimize fatigue are very promising.

21 ACKNOWLEDGEMENTS ill The authors are inebte to the University of California for it's support. We thank Mr. John RarMling for his avice an assistance uring the calibration an software evelopment phases of the project. Finally, we are grateful to Ms. Lesa Havert who very ably prepare the manuscript. RE FERENCES Houtz, S.F., an Fisher, F.J., (1959) "An Analysis of Muscle Action an Joint Excursion on a Stationary Bicycle," The Journal of Bone an Joint Surgery, Vol. 41-A, No. 1, pp Hull, M.L. an Davis, R.R., (1981) "Measurement of Peal Loaing During Bicycling: 1. Instrumentation," Journal of Biomechanics, Vol. 14, No. 12, pp Tate, J. an Shierman, G., (1977) "Toe Clips: Efficiency," Bicycling, Vol. 18, No. 6, p. 57. How They Increase Pealling Whitt, F.R. an Wilson, D.G., (1974) Bicycling Science, MIT Press, Cambrige.