2.3.2. Upper Limb Muscle Load Tests (EMG)

Noraxon [12] apparatus was used to measure the electromyography signal. The EMG signal was recorded with the use of surface electrodes. This signal, which gives a picture of muscle activity, is strongly dependent on internal (individual factors) and external [13] factors. So, it is not a signal with constant parameters (such as body temperature). Internal factors include conduction velocity of muscle fibers, thickness of the subcutaneous layer (fat content), and the proportions of a specific type of muscle fibers. External factors are related to the type of measuring electrodes used and the electrical resistance between the electrode and the skin. Moreover, the EMG signal depends on the geometric factors influencing the change in the length of the muscle (here: resulting from the change of body position and the technique of driving on the tested device). Among the factors influencing the EMG signal, the physical activity of the tested persons should also be specified. The analysis of the results of the experimental tests was carried out based on the most frequently used parameters of the EMG signal. These parameters are determined with respect to time (RMS amplitude) and frequency (MPF average frequency, determined following the Fourier transform).

The first type of analysis was used in the study because the EMG signal amplitude is an important indicator of muscle activity. The analysis of the EMG signal amplitude can be performed based on the root mean square (RMS) amplitude. The method of determining the RMS parameter is described in Relation (4). The RMS parameter is determined from signal fragments (so-called windows) of a specific length.

$$RMS = \sqrt{\frac{\sum\_{i=1}^{n} X\_i^2}{n}} \tag{4}$$

where *n* is the number of samples to be analyzed (window length); *Xi* is the value of the *i*-th sample.

The basic parameters of the experiment are defined below.


The division into time intervals significantly influences the conditions of the analysis. In general, an EMG signal is not a stationary signal. It may show changes over the course of the experiment (e.g., a downward trend which is a measure of fatigue). It is assumed that the signal is locally stationary in the analyzed time intervals. The time intervals "overlap", which means that the obtained parameters (in this case, RMS) in adjacent intervals are strongly correlated with each other.

The subject of the analysis was six muscles from Table 1. The selection of muscles was made based on the literature analysis [14–21] as well as consultations with a physiotherapist [21]. Figure 5 shows the visualization of the muscles selected for the study.

**Figure 5.** Visualization of the examined muscles/own work based on anatomical atlases.


### **Table 1.** Analyzed muscles.

EMG signals were measured using surface electrodes placed on the right limb of each participant, as shown in Figure 6. The signals were recorded while performing the assumed rides ("route", "slalom").

**Figure 6.** Location of electrodes for surface EMG measurement used in the experiment.

The obtained "raw" EMG results needed to be normalized. The most common method, known as MVC normalization, refers to the maximum voluntary contraction (Xmax; shown in Table 1) prior to starting the test measurements. Typically, MVC contractions are performed against static resistance. Ordinary (untrained) participants in the experiment may have difficulty reaching the MVC level as they are not used to such an effort. For patients who cannot, and should not, perform MVC due to damaged muscle structures, other methods of processing and analysis should be considered. Ultimately, the subject of the analysis in the experiment was the signal determined by the formula:

$$Xa = \frac{X\_{measured} - X\_{minimum}}{X\_{maximum} - X\_{minimum}} \tag{5}$$

where: *Xminimum* represents EMG signal emitted at no load.

#### 2.3.3. Evaluation of the Steering Wheel's Functionality

At this stage of research, it was necessary to consider how to verify the functionality of the universal steering wheel. In this case, the functionality of the steering wheel is understood as its influence on driving correctness and level of muscle load. This problem should be worked out using methods of statistical analysis based on the methods of experimental research. The properties of correctness of driving with a standard steering wheel with pedals were adopted as a reference level of the functionality. It was so because this type of steering wheel is a long-time, well-established solution. However, it needs to be remembered that the standard steering wheel does not make it possible to drive a vehicle for people with disabilities of the lower limbs. Thus, it will not find its application in the eco-car. Thus, the aim was to find an answer to the question of whether the newly designed ECO steering wheels (intended for use in the eco-car) are better or worse than the standard steering wheel, in terms of the analyzed test drive quality indicators. In order to achieve this goal, the following evaluation algorithm was adopted:


This algorithm, sketched very generally, required the use of a variety of statistical tests such as, for example, the Shapiro–Wilk test for normality, the Friedman's test of concordance (substitute for the ANOVA test for nonparametric hypotheses), and the Wilcoxon test for the relationship of paired observations. In the case of variables dependence confirmation, a simple post hoc analysis was performed, consisting of the classification of differences in the quality indicators of paired observations.

#### **3. The Results of the Experiment**

The following graphs show the results for one tested driver. As mentioned earlier, the sample included similar people, of similar age, skills, and one gender. This was reflected in relatively small differences in the obtained results, at the level of 10%. The presented results for a specific driver are therefore representative for the entire analyzed sample.

The h parameter is the RMS value calculated for time intervals (not for the entire duration of the experiment T). The physical interpretation of the h parameter is shown in Figure 7. Figures 8–10 show the course of the h parameter as a function of time for different types of steering wheels and the "route" type ride.

**Figure 7.** Interpretation of the h parameter. The parameter is measured in the radial direction to the reference trajectory and concerns the geometric center of the vehicle.

**Figure 8.** Values of the h parameter for the "route" type ride: (**a**) NOR steering wheel: red, (**b**) ECO 180 steering wheel: green, (**c**) ECO 120 steering wheel: blue.

**Figure 9.** Values of the h parameter for the "slalom" ride: (**a**) NOR steering wheel: red, (**b**) ECO 180 steering wheel: green, (**c**) ECO 120 steering wheel: blue.

**Figure 10.** Graphs of deviations of the driving speed from the set value for three steering wheels and two types of rides: "route" and "slalom". The graphs are color-coded: red: NOR steering wheel, green: ECO 180 steering wheel, blue: ECO 120 steering wheel.

#### *3.1. Analysis of the Correctness of Performing Maneuvers*

The necessary condition for the ride to be considered correct (along with the maneuvers that everyone had to perform) is that the car does not leave the lane boundaries (one lane), and in the case of the moose test ("slalom" type ride), it does not leave the road (two lanes). This condition was met by all participants of the experiment. Then, the subject of detailed analysis was to determine the h deviation from the reference line.

The subject of comparative analyses in the group unit is the parameters of the characteristics of the runs made for three types of steering wheels: NOR, ECO 180, and ECO 120.

The charts relating to the "route" type ride show no particular differences. However, driving with a standard ratio steering wheel gives the smoothest ride. More sensitive steering wheels (ECO 180 and ECO 120) show a much greater "yawing" effect. The driver has more difficulty keeping the rectilinear motion. This effect is very clear and as expected. A similar effect also occurs during the "slalom" test ride. This situation is even more surprising since for more sensitive steering wheels it is reasonable to expect an increase in precision of performing a turning maneuver. However, also in this case, a deterioration of the values of the quality indicators describing the ability to follow the reference trajectory was observed. This effect may be related to the acquired habits of drivers resulting from the practice of driving cars with the standard steering wheel. It is difficult for the driver to adapt the correct angle of steering wheel rotation to the roadway's turn (curve) and speed of the car. This effect might not have been observed if drivers had received additional training with more responsive steering wheels. Figure 10 shows the course of the deviation of the speed from the set speed (changed while driving) for the "slalom" ride. For the "route" type ride, the entire ride was performed at a constant speed of V = 50 km/h. For the "slalom" ride, the driver had to follow the speed value displayed on the screen, varying in the range of 20–50 km/h.

While analyzing the obtained experimental waveforms of the lateral deviation from the track axis and deviations from the required driving speed of one of the drivers, a preliminary assessment was made. However, an unequivocal answer can only be obtained from the results of statistical analyses of experimental studies with the use of appropriate confirmatory—confirming or confirmatory—rejecting procedures described in our algorithm. So, having conducted the said analyses, it was found that eco steering wheels in the "route" test have poorer quality indicators than the standard steering wheel: they increase the yawing effect, the effect of uneven driving speed, and the effect of lateral distance from the designated drive line. However, no case of falling out of the permitted

lane was registered for any of the steering wheels. Figure 11 shows the results of post hoc analysis for the value of the lateral deviation from the designated drive line. The differences in the number of paired observations are presented. The results are marked as follows: N: observations assigned to the standard steering wheel, 120: observations assigned to the ECO 120 steering wheel, 180: observations assigned to the ECO 180 steering wheel, 1: difference class N-120, 2: difference class N-180, 3: difference class 120–180. Dark bars indicate a positive result, and lighter ones a negative result. Since the higher value of the indicator constituting the observations determines the deterioration of the ride quality, the obtained result confirms a significantly lower functionality of the ECO steering wheels as compared to the standard steering wheel (for the indicator in question).

**Figure 11.** Graph of the number of observation classes for paired values of the effective transverse distance from the designated route line.

In the case of the "slalom" test, no statistically significant differences were found between the distributions of random variables of the paired observations. This result means that there is no reason to reject the hypothesis that the steering wheel functionality is identical.

#### *3.2. Analysis of the Results Obtained from EMG Measurements*

As before, the results for one tested driver are presented. Figure 12 shows the graphs of the EMG RMS signal for the NOR steering wheel, Figure 13 for the ECO 120 steering wheel, and Figure 14 for the ECO 180.

In the case of the standard NOR steering wheel, the most active muscles are PROS and ZGIN, i.e., the extensor and flexor of the fingers. Their activity is related to gripping the steering wheel.

The results presented in Figures 12–14 show, first of all, differences in the muscle load for particular types of steering wheels. For the standard steering wheel, for example, a smaller load on the first interosseous muscle can be observed, as compared to other types of steering wheels. In the case of ECO 120, we observe a statistically reliable increased activity of the flexor muscle of the fingers (in relation to the standard steering wheel). The greatest load on the wrist flexor is for the ECO 120 steering wheel, which is probably due to the so-called phenomenon of "yawing", i.e., the need to correct the vehicle movement due to excessive sensitivity of the steering system. A similar conclusion can be drawn for the pectoralis major muscle. The described conclusions were repeatable for approx. 80% of the analyzed sample of drivers. In the case of 20%, there were differences, which, however, did not have a constant trend. No physical fatigue of the muscles was observed in any of the cases, which would be reflected in the RMS signal trend. Perhaps it was also due to the relatively short duration of the experiment. The subjective assessment of the comfort of using a given type of steering wheel was assessed with the use of questionnaire surveys. The participants of the experiment had to indicate which steering wheel suited them best. The results were quite interesting: 57% of the participants indicated a steering wheel with a standard gear ratio, 23% a steering wheel with an ECO 120 gear ratio, and

20% a steering wheel with an ECO 180 gear ratio. In the case of people with disabilities, other results can be expected due to the frequent correlations between the impairment of the lower locomotor system and the partial impairment of the upper locomotor system, which prevent the use of a standard steering wheel (with pedals). Different preferences of the respondents in the choice of steering wheel prove that there is a need to adjust steering wheel parameters to the individual expectations of drivers.

**Figure 12.** Graphs of the EMG RMS signal for the NOR steering wheel: (**a**) "route" type ride, (**b**) "slalom" type ride. The order of the graphs corresponds to the following muscles: MIE: "interosseous", PIER: "pectoralis", NARA: "deltoideus", PROS: "extensor", ULNA: "ulnaris", ZGIN: "flexor".

**Figure 13.** Graphs of the EMG RMS signal for the ECO180 steering wheel: (**a**) "route" type ride, (**b**) "slalom" type ride. The order of the graphs corresponds to the following muscles: MIE: "interosseous", PIER: "pectoralis", NARA: "deltoideus", PROS: "extensor", ULNA: "ulnaris", ZGIN: "flexor".

**Figure 14.** Graphs of the EMG RMS signal for the ECO120 steering wheel: (**a**) "route" type ride, (**b**) "slalom" type ride. The order of the graphs corresponds to the following muscles: MIE: "interosseous", PIER: "pectoralis", NARA: "deltoideus", PROS: "extensor", ULNA: "ulnaris", ZGIN: "flexor".

However, a complete answer to the question about the possible change in muscle loads during the use of different types of steering wheels can only be obtained from a statistical analysis. In this case, its limitations should be described. First of all, the number of drivers for whom EMG tests were performed is much smaller than the number of drivers who performed simulator rides. This fact translates into a significant reduction in the test power in the case of the application of the confirmatory—rejecting procedure. In addition, the rides made were short (a few minutes), which turned out to be too short for the rides to generate significant muscle fatigue. For this reason, the collected data turned out to be insufficient to perform a full analysis. Only in the case of the study of the differences in the EMG RMS trend of the pectoral muscle (pectoralis) was there a statistically reliable result, confirming greater muscle effort when driving the car with ECO steering wheels. This situation is illustrated in Figure 15.

**Figure 15.** Number of observation classes of paired EMG RMS trend values of the pectoral muscle in the "route" test.
