1. Introduction
During ski jump landing preparation, as well as during the entire flight phase, ski position plays an important role in performance and safety. The ski position during landing preparation has been shown to increase the jump length by up to three meters [
1]. In fact, a larger angle of attack (i.e., the angle between the ski and the air stream) enables the ski jumper to exploit the aerodynamic lift force and its cushioning effect. This effect permits the athlete to decelerate, with a consequent reduction of the impact forces [
1] and, consequently, of the injury risk [
2]. Delaying the landing preparation time has been demonstrated to be one of the performance factors, together with an effective take-off, a high initial velocity, and an efficient flying technique [
3], and what distinguishes the high-ranked jumpers from the low-ranked ones [
4]. The start and duration of the landing preparation have not been defined yet, but major differences of the ankle and hip angles were observed at 0.4 s before the landing impact, while knee joint variations were found at 0.14 s before the landing [
4]. Moreover, during competitions, the landing technique is evaluated according to the Competition Rules of the International Ski Federation [
5], and constitutes part of the points of the final score, together with jump length, wind factor, flight technique, and starting gate [
5]. In particular, the athlete should land using the so-called “telemark”, a step landing position, difficult to perform but biomechanically more advantageous than the parallel leg landing (i.e., landing with the feet at the same height in a squat position) [
2]. As a consequence, a correct ski positioning and timing of the start of the landing preparation permit the athlete to execute a correct telemark position, as well as reduce the impact force acting on the lower limbs [
2].
During the flight phase, having a stable ski position is essential for performance and safety [
3,
6,
7]. The ski jumper usually tries to keep a V-style ski position, since it has been shown to be more effective than the parallel position [
6]. To achieve aerodynamic efficiency and stability, the athlete needs to continuously adjust his/her ski movements in order to compensate for external factors (such as the change of pressure and wind) that are acting on him/her, finding a compromise between a steady position and angular adjustments [
7].
Consequently, the goal of the present study was to investigate the ski position during the ski jumping performance due to the aforementioned important role played during the flight and the landing preparation phases. The detection of the ski orientation could support trainers and athletes in improving technique and performance. The ski opening angle and the movement regulating the angle of attack have been determined using 3D video analysis. However, the rotation around the longitudinal axis of the ski (roll), responsible for the tilting movement, appeared to be inaccurate using video cameras, due to the difficulties of visually determining the ski rotation [
8]. Compared to the 3D video analysis, the use of wearable sensors, such as inertial motion units (IMUs), could constitute an interesting set-up for in-field ski movement analysis, being able to detect the orientation of the skis more accurately.
IMUs placed on skis have been applied in previous studies dealing with skiing sports, in particular, cross-country (XC) skiing, ski mountaineering, and ski jumping. In XC skiing [
9], a fixed sensor on the ski has been used to determined cycle duration, speed, and distance. In addition to these variables, skin off and on, kick-turns, slope angle, and elevation gain have been detected by IMUs in ski mountaineering [
10,
11]. With or without further sensors on body segments, IMUs placed on the skis have been used to detect the sub-techniques of classic [
12,
13] and skating [
12,
14] XC skiing techniques. Moreover, IMUs on the skis have been used to analyze the friction between skis and snow [
15]. Finally, in ski jumping, inertial sensors have been employed by different authors, being light and with a wide recording volume [
16], two important characteristics of wearable sensors for their use in this sport. Previous publications carried out data collections using IMUs on skis and on body segments in order to analyze the overall performance [
17,
18,
19], or the take-off and in-run [
20], or the lower body kinematics during the landing impact [
21], but without deeply concentrating on the ski angular movement. The potential of the use of inertial sensors placed only on the skis to detect their position has been introduced and tested on one subject by Kreibich and colleagues [
8]. Always with the sensors placed on the skis, Groh and colleagues [
22] were able to detect the ski speed and the jump length. Moreover, the same author introduced the use of inertial sensors on the skis to detect the angular momentum during landing, validating it with custom-made force-measuring bindings [
23].
To the best of our knowledge, no studies investigated the skis’ movement during landing preparation and the possible correlations with the landing kinetics, important factors for injury prevention and performance improvement. Therefore, a combination of IMUs placed on the skis and wireless force insoles could represent an interesting set-up for this analysis. This combination has been previously introduced by the authors of the present paper, and the first results showed that the ski position influences the vertical ground reaction force (GRF) [
24]. The combination of IMUs and force insoles proposed in [
24] was utilized in the present study on a higher number of ski jumpers to detect possible correlation between ski position and impact kinetics. Moreover, an overview of the ski movements during the flight phase was presented.
The goal of the study was to achieve greater insight into the ski position during the flight phase by means of inertial sensors, with a particular focus on the landing preparation in order to detect correlations with the impact kinetics. We hypothesize that:
- (i)
Each athlete owns his specific ski pattern during the flight performance, depending on the competition level and expertise [
7];
- (ii)
The pitch (rotation around the frontal axis) during the landing preparation is the ski movement that mainly acts on the impact kinetics, being related to the angle of attack [
1];
- (iii)
The roll (rotation around the sagittal axis) during the impact influences GRF, since it influences the direction of GRF resultant vector;
- (iv)
Around 0.40 s before the landing impact as in [
4], the main ski movements that lead to the start of the landing preparation happen.
4. Discussion
Referring to our hypotheses, the results showed that: (i) each athlete owns his specific ski pattern during the flight phase; (ii) the pitch during the landing preparation is the ski movement that mainly acts on the impact kinetics; and (iv) significant ski pitch variations happened between t0.36 and t0.16, leading to the consideration that around 0.36 s before the landing there is the start of the landing preparation. However, different from the hypothesis (iii), the roll during ts did not correlate with any of the kinetic variables during landing impact.
4.1. Ski Movement during the Flight Phase
As visually notable from the reported cases (
Figure 4 and
Figure 5), the curves of the ski angles are distinctive among the participants, owning their personal movement patterns depending on the expertise of the athlete [
7].
During the phase of the first flight (i.e., the transition phase between the end of the take-off and the start of stable flight), the athlete needs to open the ski in a V-shape to rotate the ski internally and to raise the ski tips in order to increase the aerodynamic force acting on himself (
Figure 4 and
Figure 5). Each ski jumper attained stable flight in a different way. For example, subjects
a,
d,
e,
f, and
j had a steep and symmetric pitch angle in comparison to the other subjects during the first flight (from 0% to 20% circa of the flight time in
Figure 4 and
Figure 5). After the first flight phase, some athletes kept a wide angle of attack (corresponding to a high angular value of ski pitch), reducing smoothly the angle during the flight phase (as
b and
f). Other athletes brusquely moved the ski in a horizontal position (around 0°), as athlete
e (
Figure 5). Moreover, some athletes (
c in
Figure 4 and
j in
Figure 5) showed an asymmetrical yaw angle.
During the flight phase (between 20% and 90% circa of the flight time in
Figure 4 and
Figure 5), the athlete should keep a stable and symmetrical position [
5,
6,
7]. Referring to
Figure 4 and
Figure 5, it is possible to notice how some athletes kept an unstable position, with a lot of adjustment during the flight phase, as, for instance, subjects
f,
h, and
j (
Figure 5). At the same time, subjects
f and
h, together with subject
i, never kept the same pitch angle during the flight, decreasing it constantly during the entire phase (
Figure 5). The ski opening angle ranged between 30° and 40°, as previously reported to be the most efficient angular position [
7,
8]. The roll angle differed among athletes: Some participants kept the ski rotated around 50° (as
e and
j,
Figure 5), while athlete
h maintained relatively flat skis during the flight phase. A flat V-style has been shown to have better aerodynamic characteristics in comparison to a V-style where the skis are not so close to the body [
32].
During the landing preparation (at the end of the flight phase, from 90% until the end of the flight time circa in
Figure 4 and
Figure 5), subject
a changed the pitch angle with rapid movements, while subjects
b,
c, and
d prepared the landing in a smoother way, changing the ski pitch slower (
Figure 4). The athletes demonstrated having generally asymmetrical ski movements, independently of the roll, pitch, or yaw angle. The asymmetry could be explained by the expertise of the athlete [
7], but also by the inconstant lift and drag forces acting during the flight that, differently from wind tunnel testing, cannot be excluded during in-field tests [
8].
Considering as criteria for judging the quality of the ski position technique the previously mentioned statement that the athlete should keep a stable and symmetrical position during the flight [
5,
6,
7], none of the athletes of the study showed an outstanding ski position technique. This could be explained by the fact that the participants of the study, despite being elite athlete members of the German National Team, belonged to the Junior category, in which a technical maturity is still not reached.
The representations in
Figure 4 and
Figure 5 are based on the normalization of ski movement data during the aerial phase of the ski jumping performance (from the end of the take-off until the landing impact). However, if the first flight and landing preparation time have a comparable duration among jumps of the same athletes, the flight phase has a different duration. This means that with a normalization to 100 samples, the data can be “stretched” or “compressed”, and therefore, potentially influence the visual representation. However, for each athlete, the three
tflight were comparable in duration; on average, in fact, the difference among the jumps was of 0.10 s (3% of the average
tflight).
Finally, the analysis of the ski movement pattern is important during daily training of the athletes, and the use of inertial sensors could replace video cameras, providing reliable data without needing a lot of time for postprocessing or placing the cameras around the ski jumping hill.
4.2. Ski Movement during the Landing and Its Preparation Phases and Influence with the Kinetics
The pitch was the main ski movement to correlate with
tflight (
Table 2 and
Table 3), confirming the role of this movement during the flight phase, given its relation with the angle of attack and the consequent influence on the aerodynamic forces. In particular, wider ranges of motion of the pitch corresponded to longer
tflight and consequently longer jumps. This means that the wider the difference between the ski pitch at landing and during the flight, the longer the jump length reached by the athlete. As a consequence, since the ski jumper needs to keep the skis flexed as long as possible in order to profit from the aerodynamic lift force [
1], the athlete has to perform the landing preparation in a short time. No correlations were found between the pitch ROM and
tflight at
t0.16. This could be related to the fact that the athletes are close to the ground of the landing area at 0.16 s before the impact. Consequently, the angle of attack, controlled by the pitch movement, cannot influence the aerodynamic forces when the athlete is too close to the ground [
1].
The magnitudes of the collected kinetic variables were comparable with the ones of previous publications [
21,
24,
33] (
Table 1). The correlation between the normal GRF
max and the ROM of the pitch before the landing (landing preparation) confirmed that the ski movements during this phase play an important role not only for the jumping performance, but also for safety, acting on the aerodynamic lift forces and their cushioning effect, reducing the impact forces [
1] and, consequently, the injury risk [
2] (
Table 2 and
Table 3). In particular, wider ROM of the pitch corresponded to smaller normal GRF
max (and impulse), while the roll and yaw did not have any correlations with the kinetic variables.
Some of the collected kinetic variables correlated with certain kinematic variables differently among the ski jumping hills. For example, the impulse and normal GRF
max acting on the left side correlated with many kinematic variables collected in Oberhof (
Table 1), but not in Ramsau-am-Dachstein (
Table 2). This could be related to the fact that different athletes carried out the data collection on the two ski jumping hills. Consequently, their personal ski movement pattern could have influenced the kinetics in a different way. Therefore, a deeper analysis of the ski position pattern, as the one previously proposed, could give further information about the relation between ski movements and landing kinetics.
Focusing on the ski pitch between t
1.00 and t
landing (
Figure 6), it is possible to notice objectively how between 1.00 s and 0.56 s before the landing, the athletes kept a stable position. In fact, the average differences between the ski position of the left and right ski in the ranges
t1.00–
t0.76 and
t0.76–
t0.56 were of 1.0° ± 4.1° and 2.1° ± 2.6°, respectively. Moreover, no significant difference was found between the variations
t1.00–
t0.76 and
t0.76–
t0.56 (
Figure 7). Therefore, due to the limited ranges of variation, it is possible to consider that the ski angular movements happening until 0.56 s before the landing are only adjustments for keeping the flying position stable. Therefore, in this phase, the athlete needs to adapt the ski movements to the aerodynamic changes he is subjected to. Between
t0.56 and
t0.36, and
t0.36–
t0.16, the pitch movements varied by 4.4° ± 3.2° and 8.2° ± 4.8°, respectively. In particular, the angular difference of 8.2° ± 4.8° between the pitch recorded at
t0.36 and
t0.16 could be considered remarkable and related to the start of the landing preparation, considering that the angular adjustments were too wide to be related only to adaptations to the aerodynamic changes. Therefore, in line with a previous publication [
4], the start of the landing preparation can be considered to happen around 0.4 s before the landing, when major movements of the hip and ankle joints were detected [
4].
Finally, it is important to keep in mind that we calculated the ski angular range of motion (ROM) between the landing impact
ts (set as reference) and specific timing before it. These timings (0.76 s, 0.56 s, 0.36 s, and 0.16 s) were chosen based on a previous publication [
4]. It can be speculated that changing the timing during which the ROM of the ski movements was calculated would also change the possible correlations with impact kinetics. However, due to the variability of the ski pattern movement among athletes, defined timing before the landing was used instead of kinematic variables.
4.3. Limitations and Methodological Considerations
A remarkable aspect of the study was that it was conducted on a homogeneous group of elite athletes competing at International level with ages ranging between 16 and 19 years old. The tests were performed on behalf of a scientific support for the Ski Federation during training camps. Despite the small number of tested subjects (10), the group represented the totality of the German Junior National Team. Therefore, due to the limited number of athletes belonging to the team, including in the data collection a higher number of subjects with the same technical abilities and experience was not possible.
A limitation of the study was that the tests were carried out in two different locations, but where the ski jumping hills had a comparable size (both K-points set at 90 m) and comparable weather conditions (sunny, no wind). Moreover, two different subgroups of athletes belonging to the National Team performed the tests on the two ski jumping hills. The reason was that during the planned data collection performed within a training camp on the ski jumping hill of Oberhof, we could collect only six ski jumpers during the first day of measurements. During the second day, in fact, due to the rain and the wind, we could not carry out the tests with the remaining part of the team, because it was not possible to guarantee the same testing conditions. Therefore, we collected the data of the other four members of the National Team during the following training camp in Ramsau-am-Dachstein, on a ski jumping hill with a comparable size, always using the same combination of IMUs and force insoles. In this way, it was possible to provide the aforementioned biomechanical feedback to all the athletes of the National Team. For clarity, in the Discussion, we concentrated only on the biomechanical variables that were statistically significant on both the ski jumping hills.
Regarding the set-up, one of the main advantages was that it was not necessary to perform a calibration of the inertial sensors before doing the data collection. In fact, thanks to the algorithm proposed by Fang and colleagues [
30], during the postprocessing, the raw data of the inertial sensors were reconstructed based on the design of the in-run of the ski jumping hill. The advantage of the post-initialization is very important, making the set-up easy to use, in case athletes and coaches would be interested in using the system on their own as feedback during training. In fact, not being professional researchers, they could introduce errors during the data collection. In addition, the combination of inertial sensors and force insoles can be considered relatively light (0.3 kg). Generally, the weight of the technological equipment used in the protocol is of significant importance when performing biomechanical research in sports, and it is essential in ski jumping, a sport in which the weight of the system equipment plus athlete is the main performance factor [
34,
35].
The low sampling rate of the loadsol insoles (100 Hz) could have affected the capability of measuring impact. However, publications related to this topic are discordant: Peebles and colleagues [
26] highlighted under-/overestimation bias of the impact force peaks when using loadsol at 100 Hz. Other research groups did not report limitations related to the sample rate [
25,
27]. At the time of the data collection, loadsol insoles sampling at 200 Hz were still not available on the market. Anyway, for further studies, force insoles sampling at 200 Hz are recommendable to improve the accuracy.
A high number of external factors (such as wind and air pressure) generally interfere with the movement of the ski jumper. Consequently, we can speculate that each jump can be considered as a specific case, also when performed by the same athlete and even though the athletes belonged to an elite level. Therefore, performing statistics is very difficult in this kind of analysis, especially when dealing with the landing that is the phase at the end of the performance and, consequently, a resultant of the previous ones [
32]. As a result, the statistics performed in this study, and generally in in-field ski jumping research, need to be evaluated carefully.
5. Conclusions
The pitch was the main ski movement influencing the magnitude of the normal ground reaction force (GRFmax) and the jump performance (tflight) due to its relation with the angle of attack. As a result, in order to increase the jump length and reduce the impact forces, the athlete should keep the ski more flexed during the landing preparation phase. The pitch started to considerably vary between 0.36 s and 0.16 s before the landing impact, leading to the consideration that the landing preparation started around 0.36 s before the impact.
Despite the elite level of the athletes, each subject showed an individually unique ski movement pattern during the flight phase. The analysis of the ski position could permit improving the aerodynamics of the athlete during the flight, since previous publications gave suggestions on the best ski configuration to increase the performance [
1,
3,
6,
7,
32,
36,
37]. However, a previous study performed in a wind tunnel showed how the aerodynamics of an isolated ski depend on the combination of the roll, pitch, and yaw angle [
32]. Therefore, further studies should focus on analyzing the combination of the roll, pitch, and yaw movement during in-field performance.
According to the feelings of the jumpers, the set-up constituted by the force insoles and the IMUs resulted in not interfering with the performance. Therefore, under the practical point of view, the already proven advantages of the IMUs [
15,
17,
18,
20] and the force insoles [
25,
26,
27,
28], as well as the possible advantages of their combination shown in the present study, could provide a reliable and objective feedback for coaches and athletes for monitoring the kinetics and kinematics of the ski jumping performance. To confirm this, a report with graphs about ski pitch and roll movements and the kinetics during the whole performance were provided to athletes and coaches at the end of each day of data collection. Both athletes and coaches provided a positive feedback about the report.