INTRODUCTION AND PURPOSE OF THE STUDY

More specifically, it was investigated whether the use of clip-on handlebars significantly altered the time taken to reach brake levers compared to times obtained with traditional hand postures on the handlebar.

The underlying assumption motivating the study, is that riding postures providing quick access to brake levers, allow the athletes to more easily manage the braking operation, when he is forced to suddenly decrease bicycle speed or change direction for avoiding an unexpected obstacle.

For this reason, an additional goal of this research was to evaluate whether the time spent in reaching the brakes, regardless of the riding position, is a factor of great or minor importance in influencing the distance needed for a bicycle to be brought to a stop in an emergency.

For the braking action monitoring, it was decided for the use of the ELITE optoelectronic motion capture system, which has gained growing popularity in the international market for its unique powerful FPSR (Fast Processor for shape recognition) image processor. The high system accuracy, precision, and acquisition speed, even in difficult experimental conditions, are proved by more than 250 scientific communications in various research areas, and were considered adequate for the purpose of this study.

Beside the accurate measures of action times, the used methodology, by monitoring the posture assumed by all upper limb segments during the hand travel to the brake lever, allowed the collection of data that could be used to further the analysis of the motor strategies adopted by cyclists in performing this movement.

Experimental protocol

Three high level road cyclists were the subjects of this study. Each athlete was tested while pedalling on his own racing bicycle mounted on an air braked roller simulator. The bikes were equipped with a Cinelli Spinaci clip-on handlebar.

While pedalling, the athletes were asked to reach, as quickly as possible, the brake levers (as they should brake their bike in the shortest time as possible) from the following starting pre-defined handlebar hand postures:

1. on the lower ends of the handlebar bend (on the drops) (Fig. 1)
2. on the lateral side of the handlebar bend (Fig. 2)
3. on the upper lateral side, with the hands around the top of the brake lever mounts (Fig. 3)
4. the most comfortable top-bar position for each athletes (Fig. 4)
5. on the middle section of the top handlebar bend with the hand spaced (Fig. 5).
6. on the clip-on handlebar (Fig. 6)
7. on the top of the handlebar bend with the hands close to the stem (Fig. 7)

Data recording equipment

At a sampling rate of 100 Hz, the ELITE motion analysis system was used to collect the 3-D kinematic variables (displacements, velocities and accelerations) of the right upper limb segments respect to the handlebar and right brake lever frame.

To this end, the configuration of the ELITE system was the following: two TV cameras paired off on the right side of the subject, while the calibrated volume for the 3-D coordinates computing was 1.25 long, 1.25 high and 0.5 wide.

Five small retroreflective markers (8 mm in diameter) were glued on the subject skin in correspondence of the following anatomical repere points (Fig. 8):

m_b1 the lateral side of the arm at deltoid insertion

m_b2 lateral ephicondyle

m_b3 distal end of the radio-ulnar joint

m_b4 distal end of the 3rd metacarpal

m₁ distal end of the 3rd phalanx

Two additional markers were located on a stick rigidly fixed to the handlebar (m₀) and on a stick rigidly fixed to right brake lever (m₂)

To verify the measurement accuracy of the system, a test was carried out before each experimental session: a stick with two spherical markers fixed on its extremities at the distance of 400 mm were moved along the whole field of view. The mean differences between the measured and actual distance of the markers fixed on the rigid bar was within 0.4 mm, in agreement to the values declared by the manufacturer.

Data elaboration

Data processing included a 2-D tracking of the markers detected by each TV camera and a 3-D reconstruction of markers seen by the two TV cameras Filtering of 3-D markers coordinates and their derivatives computing were performed by using the algorithms developed by D'Amico and Ferrigno (1990). The algorithms are based on an autoregressive model, fitted to the signal, to evaluate the filter bandwidth and the extrapolation of the data. Then, the components of the coordinates of each marker are filtered by a linear phase FIR low-pass filter, with a proper cut-off frequency depending on the frequency content of the signal.

To account for the bicycle movements, the measured coordinates were then referred by trigonometry to a reference system fixed to the bike handlebar (Figure 11)

As showed in Figure 10, the time the rider takes to reach the right brake lever was defined as the time from the first detectable movement of the marker m₁ until the first occurrence of the m₂ marker movement. In the trials triggered by an external stimulus the pre-motor time was also measured as the time from the signal presentation until marker m₁ moves.

Statistical analysis

The effect of the different kinds of handlebar hand position on brake lever reaching times was analyzed with one-way ANOVA. The existence of significant between-condition differences was then tested using the Newman Keuls post hoc test. To determine whether the brake lever access time with the clip-on handlebar position was different from each of the other conditions, Student's t-test for paired data was also performed. All comparisons were considered significant at an alpha level of P<0.05.

Figure 12 illustrates, as a function of the different riding positions, the mean brake levers reaching-times (in ms) produced by all the subjects in the different experimental variants. Data were obtained analyzing all the 630 trials acquired.

The values ranged from 147 ms of posture 3 (hands on the upper lateral side of the bend, close to the top of the brake lever mounts) to the 276 ms of position 7 (hands on the top of the handlebar bend, close to the stem), with a largest difference of 129 ms.

According to the found mean duration values (always less than 300 ms) the examined movement could be classified as a short, pre-programmed motor responses, controlled via an open-loop process, without any possibility to change the movement pattern once started.

As expected, postures with hands placed around the brake levers were characterized by slightly lower brake access times compared to the others. In concert with some previous studies on human movement, these results indicate that the movement time, leaving aside other factors (e.g. movement amplitude and trajectory, number of joint involved etc.), heavily relies on the spatial distance from the starting and the final position.

No differences were found between the position with the hands on the clip-on handlebar and position 5 (hands on the middle section of the top handlebar bend) and 7 (hands on the top of the handlebar bend in proximity of the stem).

Estimate of the effects of the time taken to reach brake levers on the total braking distance. A comparison between the riding positions characterized by the minimum and maximum brake lever reaching-times

One of the goal of this study was to evaluate whether the time spent in reaching the brakes, regardless of the riding position, is a factor of great or minor importance in influencing the distance necessary for a bicycle to be brought to a stop in an emergency.

To this end, the following simple calculations were performed:

First, the bicycle travel distances, for three typical riding speeds (36, 48 e 60 km/h), during the time it took for brake levers to be reached, are computed and plotted in Figure 13. The results, obtained from the time values showed in Figure 12, revealed a maximum difference among the analyzed postures of 1.3, 1.7, and 2.1 m at 36, 48, and 60 Km/h respectively.

Then, these extreme values were related to the total braking distance. Assuming a maximum braking deceleration of 6m/s² (hard braking without causing loss of control), the distance the bike travels before coming to rest from when the athlete applies the brakes, was determined for the three chosen speeds.

Finally, the distance the bike will travel during a pre-motor time of about 180 ms (typical value in laboratory tasks) and a brake lever reaching-time of 260 ms was added. In doing so, the total braking distance from the instant the rider perceives the danger was estimated. The values were found to be 13, 21, and 30 m at 36, 48 and 60 km/h. These results would indicate that the maximum riding posture induced differences account for a percentage of 10, 8, and 7% of the total braking distance, assuming a bike speed of 36, 48 and 60 km/h, respectively (see Figure 14).

However, it should be noted that the pre-motor time value used for the calculation (180 ms) refers to a typical laboratory task with the subject knowing that a trigger signal it is going to be presented. In real world situations, with the subject required to react to an unforeseeable stimulus, longer pre-motor times up to 600 ms were found. Given the above, the calculated percentage values may be even overestimated.

In summary, the results here presented indicate that:

in each of the examined experimental conditions, the time the athlete takes to reach the brake levers (lower than 300 ms) represents a small fraction of the total time necessary to stop a bicycle after a emergency signal was perceived.

hand postures slightly affect the access time to the brake levers. However, this is true only involving in the comparisons postures where the hands are placed very close to the brake lever mounts. Anyway, it was demonstrated that, even in the worst case, the found posture-induced differences account for less than 10% of total distance the bike travels before coming to rest.

riding with the hands placed on the clip-on handlebar leads to brake lever reaching-times as much as those measured adopting others traditional postures with the hand placed on the handlebar.