Abstract
Objective: The objective of this study was to examine the influence of bicycle design and speed on the head impact when suffering from a single-bicycle crash, and the possibility to study this using crash tests.
Methods: Simulations of single-bicycle crashes were performed in the VTI crash safety laboratory. Two bicycle crash scenarios were simulated: “a sudden stop” and “sideways dislocation of the front wheel”; using four different bicycle types: a “lady’s bicycle”, a commuter bicycle, a recumbent bicycle and a pedelec; at two speeds: 15 and 25 km/h. In addition, sideway falls were performed with the bicycles standing still. All tests were done with a Hybrid II 50th percentile crash test dummy placed in the saddle of the bicycles, with acceleration measurements in the head.
Results: The crash tests showed that a sudden stop, e.g. a stick or bag in the front wheel, will result in a falling motion over the handle bars causing a forceful head impact while a sideways dislocation of the front wheel will result in a falling motion to the side causing a more moderate head impact. The falling motion varies between the different bicycle types depending on crash test scenario and speed. The pedelec had a clearly different falling motion from the other bicycle types, especially at a sudden stop.
Conclusions: The study implies that it is possible to examine single-bicycle crashes using crash tests, even though the setup is sensitive to minor input differences and the random variation in the resulting head impact values can be large. Sideway falls with the bicycles standing still were easier to perform with a good repeatability and indicated an influence of seating height on the head impact.
Introduction
Swedish government advocate for increased cycling and, at the same time, a decrease in injury rates among cyclists (Government Offices of Sweden Citation2017). According to accident data from emergency hospitals, eight out of ten severely injured cyclists in Sweden have been involved in a single-bicycle crash (Niska and Eriksson Citation2013). An international comparison shows that single-bicycle crashes represent 60 to 90 percent of cyclists admitted to hospitals (Schepers et al. Citation2015). Nearly half of the single-bicycle crashes in Sweden are related to maintenance and operation, with the main cause being a loss of friction (Niska and Eriksson Citation2013). Causes related to the road design, the cyclist’s handling of the bicycle, the cyclist’s behavior and condition account for roughly 15 percent each, while one out of ten is related to the interaction with other road users. Many of these single-bicycle crashes are a result of a “sudden stop” induced by factors such as curb stones/sharp edges (11%), potholes or other road surface defects (8%), temporary (3%) or permanent objects (7%) on the road or bicycle path, rails (2%), a bag/stick/etc. getting caught in the bicycle (6%), handbrakes locking the front wheel (5%), dogs or other animals (3%). More than half (51%) of the severe injuries from a single-bicycle crash are to the arms and shoulders. Legs including hips account for 23 percent of the injuries and the head and face account for 8 percent (Niska and Eriksson Citation2013).
To get a better understanding of the mechanisms behind single-bicycle crashes and complement numerical accident reconstructions in computer simulations (Bauer et al. Citation2016; Fahlstedt Citation2015), simulations in crash test laboratories could be a possible method. Crash tests including bicycles are scarce and when performed they usually simulate collisions with motor vehicles (Watson Citation2010), not single-bicycle crashes.
The objective of the study presented in this paper was to examine the influence of bicycle design and speed on the head impact when suffering from a single-bicycle crash, and the possibility to study this using crash tests.
Methods
Single-bicycle crashes were simulated in the indoor facility of the VTI crash safety laboratory using a Hybrid II 50th percentile crash test dummy. Two bicycle crash scenarios were simulated:
A sudden stop - representing “a stick/a bag in the front wheel”, “a sudden lock of hand brakes” or “hitting a firm object”.
A sideways dislocation of the front wheel – representing “loss of friction” or, “obliquely hitting a curb stone or another firm object”.
To perform the simulations, a specially designed rig was constructed and mounted on top of a sledge normally used when crash testing child restraint systems. That rig kept the crash test dummy and bicycle in position during the propulsion phase when the system was accelerated to a desired speed. When braking the sledge, the bicycle with the crash test dummy in the saddle continued to roll freely forward from the rig, at the desired speed before the “crash”. To simulate crash test scenario 1, a metal hook was mounted on the front wheel. When the wheel had turned three-quarter of a revolution, the hook was stuck in the fork of the bicycle and a sudden stop occurred. To simulate crash test scenario 2, a square steel tube was mounted on the floor one meter in front of the stopping point of the rig. The tube of 150*150 mm was installed obliquely so that the front wheel of the bicycle would hit the tube at a 20-degree angle. The instruments and equipment used are further described in Supplemental Material Appendix.
About 30 pretests were needed to modify the construction and finalize the test procedure to achieve reasonable accuracy and repeatability, meaning that the bicycle with the crash test dummy was “delivered” in a similar way and at the same spot in each test and that the falling motion of the crash test dummy and the bicycle appeared to be the same in consecutive crash tests with the same settings. When that was achieved the actual tests were performed using four different types of bicycles:
An open frame bicycle with an up-right seating position, i.e. a “lady’s bicycle”.
A closed frame bicycle with a more forward leaning seating position, referred to as a “commuter bicycle”.
A recumbent bicycle, with a low and backwards leaning seating position.
A closed frame pedelec, with a rather up-right seating position and the electric support engine enclosed in the front wheel hub and the battery positioned in the lower cross member of the frame.
The included bicycles varied in respect of seating height, seating position and weight distribution, with the purpose to study how that would influence the falling motion and the consequence of a crash. In the choice of bicycles, we strived to include regular bicycle types while it was important to also get a variation in constitution. The lady’s bicycle and the commuter bicycle were considered to represent some regular bicycle types. Observations on bicycle paths in Sweden indicate that these types could represent between 70 and 93 of the bicycles used in Sweden (Eriksson et al. Citation2019). Since pedelecs are becoming increasingly popular and are heavier and have a different weight distribution than a regular bicycle, such a bicycle was also included. Although relatively rare, the recumbent bicycle was included since it differs substantially from the other bicycles in terms of seating height and seating position. By including different types of bicycles, we could also evaluate the robustness of the method in the sense of different types of bicycles.
Each bicycle type and crash scenario were tested at two different speeds: 15 and 25 km/h - representing low and high average cycling speeds in Sweden (Eriksson et al. Citation2019). In addition, mainly as a reference, sideway falls were performed with the bicycles standing still (0 km/h). Each set-up was performed at least twice to study the repeatability of the method. In total, almost 40 simulations of single-bicycle crashes in motion were performed and 16 sideway falls with a bicycle standing still. Head acceleration data was recorded at 10 kHz, and to remove disturbances, flicker and extremely short peaks in the measurement, which could interfere with the interpretation of the results, data was filtered digitally using a Channel Frequency Class filter (CFC 1000). In addition to acceleration measurements in the crash test dummy head, the tests were documented with several video cameras at different angles, including high-speed video cameras. The head of the crash test dummy was smeared with a red colored paste giving colour-marks on the floor when hitting the ground. That made it possible to detect the locations of impact on the head and to measure the throwing distance of the crash test dummy from the initial point of the crash scenario to impact, and finally the sliding distance to a complete stop of the dummy.
Methods of analysis
The methods, limit values and evaluation criteria normally used in crash tests are quite specific to motor vehicle collisions (at higher speeds and with a different type of crash violence than in a bicycle crash). Since there is no “standardized bicycle crash test method”, it was not obvious how to analyze the collected information from our simulations of single-bicycle crashes.
We have chosen to present the maximum resultant of accelerations from the recordings in the crash test dummy head. To describe the crash violence represented by the acceleration values measured, we have also calculated the Head Impact Criterion, HIC36 (“36” refers to a 36-millisecond long time frame in which the resultant of the x, y and z components of the acceleration forces is integrated). HIC is a commonly used acceleration data processing procedure within the vehicle crash testing industry, describing crash violence to the head, but the measure is normally used to evaluate direct head impact inside of a car. There are no fixed threshold levels of HIC for achieving injuries or trauma, but higher values are equal to higher probability of trauma. The American NHTSA have specified a limit value of 1000 for HIC36, which indicates the level at which 50 percent of the injured will suffer permanent medical impairment (Eppinger et al. Citation1999).
In addition to analyzing the head acceleration values measured, the video recordings were carefully studied, especially the slow-motion pictures. This was done manually by two observers discussing the results together. They focused on the falling motion, for example, in what order the different “body parts” of the crash test dummy hit the floor. The purpose was partly to describe the falling motion itself and identify any variation in relation to bicycle type, accident scenario and speed, and partly to study the repeatability of consecutive tests with the same settings.
Results
Sideway falls
Sideway falls performed with the bicycles standing still showed, with a few exceptions, that the accelerations measured were consistent between consecutive tests with the same bicycle type (). A variation could be seen between the different bicycle types, indicating that a higher seating position might result in a more forceful head impact when falling with a bicycle standing still, the recumbent bicycle excepted.
Table 1. Maximum resultant of accelerations measured in the crash test dummy head when hitting the floor after sideway falls with four different types of bicycles.
The recumbent bicycle resulted in the lowest HIC-values obtained but produced a greater variation than the other bicycle types (Figure S6 in Supplemental Material Appendix). Note that the HIC-values are not always perfectly consistent with the head acceleration values recorded, since they are also taking the duration of the head impact into account. The force to the head is influenced by the falling motion of the dummy. For example, when the shoulder hits the floor before the head, the head impact is dampened. The falling motion also determines which part of the head hits the floor which, in turn, influences the acceleration values registered. According to the video recordings, the falling motion with the recumbent bicycle was not as consistent as that of the other bicycles. While the crash test dummy usually fell straight to the side at sideway falls, it fell out of the saddle during the falls with the recumbent bicycle. In two cases, this resulted in a hit to the forehead (instead of the side of the head) of the crash test dummy, which is probably the explanation for the higher values in these cases.
Single-bicycle crashes in motion
The observations of the video recordings showed a tolerable repeatability of the crash tests performed. The bicycle with the crash test dummy was “delivered” in a similar way and at the same spot in each test. The falling motion of the crash test dummy and the bicycle appeared to be the same in consecutive crash tests with the same settings – same bicycle type, speed and crash test scenario. A sudden stop resulted, in general, in a falling motion over the handle bars with the head taking the initial hit to the ground. A sideways dislocation of the front wheel resulted in a falling motion to the side with foot, knee, hip and shoulder hitting the ground before the head. Photo sequences from the crash tests are shown in Supplemental Material Appendix.
The repeatability of the simulated crashes in motion was not as high as when simulating sideway falls, resulting in a greater variation in the recorded accelerations (), both within the same settings and between different crash test scenarios and bicycle types. Small variations in steering angle or dummy seating position influenced the falling motion and thus also the acceleration values measured. For three of the crash tests performed, results from the acceleration measurements are missing. In two tests the accelerometer malfunctioned, and in one test (with the pedelec) the bicycle produced a deviating falling pattern, out of range of the measurement.
Table 2. Maximum resultant of accelerations measured in the crash test dummy head when hitting the floor after simulated single-bicycle crashes in 15 or 25 km per hour with four different types of bicycles.
Despite the random variation, it can be concluded that a sudden stop will result in a more forceful head impact than a sideways dislocation of the front wheel, except for the pedelec. Both the maximum accelerations () as well as the HIC-values () indicate this. The falling motion over the handle bars in the sudden stop scenario resulted in a distinct, forceful hit to the forehead of the crash test dummy. The falling motion to the side, in the sideways dislocation of the front wheel scenario, resulted in a more moderate head impact with a hit to the side of the head just above the “eye” or the “ear” (see photos in Supplemental Material Appendix).
Figure 1. Calculated HIC36 values from the accelerations measured in the crash test dummy head when hitting the floor after simulated single-bicycle crashes with four different types of bicycles. At the top the sudden stop scenario in 15 followed by 25 km per hour. The two lower figures show the dislocated front wheel scenario in 15 and 25 km per hour respectively. Raw data filtered using a CFC 1000 filter.
The falling motion varied somewhat between the different bicycle types. At a sudden stop, the pedelec produced a clearly different falling motion compared to the other bicycle types resulting in noticeably lower acceleration and HIC values. Instead of a falling motion over the handle bars, the pedelec resulted in a fall to the side even at a sudden stop, resembling the motion with a sideways dislocation of the front wheel. The higher weight and the lower center of gravity of the pedelec is probably the explanation for this. The recumbent bicycle was expected to generate lower acceleration values than the other bicycles, due to its low seating position. However, in one simulation of the sudden stop scenario, at 25 km/h, also the recumbent bicycle generated a forward falling motion with a forceful head impact as a result. At the lower speed (15 km/h) the recumbent bicycle did just tip over sideways.
A higher speed seems to generate higher acceleration and HIC values, although not significantly higher (p = 0.075 and p = 0.083 respectively). Especially at a sudden stop with the commuter bicycle or the recumbent bicycle, and with a sideways dislocation of the front wheel of the lady’s bicycle, the speed seemed to be of importance. In other cases, the accelerations measured in the sideways dislocation of the front wheel scenario were in line with those from sideway falls at 0 km/h. A higher speed also seemed to result in longer throwing distances of the crash test dummy after the crash scenario was induced (see Table S4 in Supplemental Material Appendix).
Discussion
An overall impression of our simulated single-bicycle crashes in motion, is that they are sensitive to minor adjustments and it is therefore difficult to achieve a high repeatability. Even minor changes in dummy placement and performance affected the free rolling behavior of the bicycle and subsequently the motion when crashing and falling. This is probably also true for real-world bicycle crashes, since even small differences in the cyclist behavior, in the crash situation, will make a difference in the accident outcome. In real crashes, several variating parameters such as speed, weight, angles, tyre-surface friction might also influence the outcome. When simulating crashes in a crash test laboratory, the repeatability is the major concern – not replicating real crashes. By trying to control all the influencing parameters – using the same speed, the same weight of the cyclist (the crash test dummy), the same crash angles and the same tyre-surface friction we have strived to produce comparable results.
The main purpose of our study was to examine the possibility of using crash tests to study single-bicycle crashes. We have found the method to be promising but the setup sensitive to minor input differences and the resulting outcome therefore somewhat coarse to evaluate, meaning that it is difficult to draw far stretched conclusions regarding details such as bicycle design. From the video recordings we could conclude that the falling motion varied somewhat between the different bicycle types which in turn resulted in different head acceleration values. From our results it is, however, not possible to draw conclusions regarding the relation between safety and bicycle design. To do that, there is a lot more to consider - for example the stability of the bicycle when riding. Nevertheless, according to our executed tests, there is no evidence that pedelecs produce an enhanced risk - which seems to be a common opinion among the public and in media. On the contrary, our crash tests indicate that the sheer extra weight and low center of gravity of the pedelec included in our tests reduced the risk of a head-on dive over the bicycle handle bars. However, that might not be the case for all pedelecs. The recumbent bicycle was expected to generate a more moderate crash violence due to its low seating position but showed, in some cases, a risky performance of tipping over the front wheel. Then again, it is difficult to draw general conclusions regarding the bicycle types tested, as there may be differences between different models and “individuals” of one and the same type of bicycle.
In the scenario of sideway falls with a bicycle standing still, we could detect a certain relationship between seating height and measured acceleration values in the head. In the simulated single-bicycle crashes in motion it was not possible to identify the same relationship. The random variation in the measured values was large and the seating height was likely to be marginal in relation to other factors. Worth mentioning is that seating height might also be a contributing factor to the crash itself, at least when it comes to cyclists over the age of 50 being involved in a single-bicycle crash (Boele-Vos et al. Citation2017).
One limitation in our tests performed is the possibility for a cyclist, in contrast to the crash test dummy, to react before the crash to keep balance, or to moderate the fall using arms or legs. Another limitation is that the crash test dummy is not developed for simulations of bicycle crashes. It is primarily designed to sit in a car and measure the forces from airbags, seat belt or dashboard in a frontal collision. With the Hybrid II-dummy that we used, it is only possible to measure the accelerations in the head. Based on accident analysis, we know that almost half of the serious injuries in bicycle accidents are arm or shoulder injuries (Niska and Eriksson Citation2013). By studying the video recordings, we have been able to get some information regarding these types of injuries. During the crash test period we also had to replace broken metal parts in the shoulder of the dummy which confirms that the impact to the shoulder can be forceful. Especially at sideway falls the shoulder is exposed to significant crash violence, and earlier studies have shown that cyclists’ shoulder injuries mainly occur at sideway falls with a straight hit to the shoulder (Stigson et al. 2014). Another crash test dummy type is available, WorldSID, with the possibility to measure forces from the side. In earlier research, when developing shoulder protection for cyclists, such a dummy has been used (Stigson et al. Citation2016). However, the WorldSID has no arms and a design that would have made it even harder to place on a bicycle. Injuries to thigh or hip are also common in single-bicycle crashes (Weijermars et al. Citation2016) and cannot be measured with the crash test dummy used. Especially among elderly cyclists these types of injuries occur, often because of falling when mounting or dismounting a bicycle (Björnstig and Näslund Citation1984; Scheiman et al. Citation2010). The simulations of sideway falls with a bicycle standing still conducted in our study, can be considered equivalent to these types of crashes. A crash test dummy specifically developed to study bicycle crashes, including sensors to measure lateral forces on arms and legs, would be valuable. Before such a dummy is available, it might be possible to mount accelerometers on hips and shoulders of the available dummies. More modern and common than the Hybrid II is the Hybrid III, but that dummy was more difficult to put on the saddle of the bicycles and more expensive to repair. As we expected the testing was tough on the dummy and we had to repair it several times during the series of testing.
The accelerometers in the head of the dummy only record straight impact forces, and not the rotation of the head caused by an oblique stroke. According to previous research, the brain is more sensitive to rotation forces than to straight hits and that can cause serious brain injuries, even though no skull fracture occurs (Gennarelli et al. Citation1972; Holbourn Citation1943; Margulies and Thibault Citation1992; Ommaya Citation1995). We tried to mount a gyro in the head of the dummy to be able to measure the rotational forces but failed to solve the technical details.
The highest acceleration values of about 910 g, which we recorded in our crash tests, are well above the 250 g limit that a bicycle helmet must be able to pass according to today's standard (Swedish Standard Institute Citation2012). This limit corresponds to a 40 percent risk of a skull fracture, but far exceeds the risk of brain injury that may already occur at 60–100 g (Zhang et al. Citation2004). In other words, the risk of skull fractures is high for the acceleration values we measured at sudden stops (344–914 g) except with the pedelec. On the other hand, the risk of skull fracture is relatively low for the acceleration values we measured in sideway falls with a bicycle standing still (40–225 g) or due to a sideways dislocation of the front wheel at 15 km/h (44–153 g), while concussion and other types of brain injuries may occur. The severity of the injury is also depending on the location of the head that is exposed to an impact, with the temporal regions being more sensitive than the forehead. A detailed study of head injuries among cyclists has shown that an impact to the temporal regions or the back of the head (occipital) was the most common among the fatal injuries (Björnstig et al. Citation1992). The head impact locations identified in our study indicate that bicycle helmets should be designed for better protection of the face and the side of the head. In standardized bicycle helmet tests, the helmet is hit right on top of the helmet and that appears not to be the most likely impact location on the head. However, our crash tests are not directly comparable to helmet tests and since the crash test dummy does not behave like a regular cyclist, the results should be interpreted with care.
Another interpretation of practical significance based on our study, is the importance of measures to prevent the sudden-stop situation for cyclists. For road authorities, it is crucial to consider the design of curb stones, attend to potholes and remove firm and temporary objects along the cycleways. Previous studies have also highlighted the importance of road design, maintenance and operation for the safety of cyclists (Niska and Eriksson Citation2013; Nyberg et al. Citation1996). For bicycle manufacturers, it is necessary to develop and distribute bicycle brakes that prevent the front wheel from locking. When it comes to the design of pedelecs it is valuable to consider the center of gravity of the bicycle, for example the placement of the battery.
Our simulations indicated that a higher speed could generate a more forceful crash violence, but the importance of speed was not significant. However, in our crash tests the impact was towards a hard and smooth surface. In reality, a cyclist could crash into firm objects at the side of the cycleway such as rocks, trees, lamp posts or parked cars and the consequence of such a crash could be severe and the speed could then be of importance. A sufficient width and lateral reserves without firm objects are important for cyclists, especially at so called cycling superhighways. From the throwing distances measured in our study, we would like to recommend a secure lateral reserve of at least two meters alongside a cycleway.
Supplemental Material
Download MS Word (7.6 MB)Acknowledgments
This study was supported financially by the research fund of the insurance company Länsförsäkringar. This support is gratefully acknowledged. Thanks also to all the staff at the VTI crash safety laboratory and at VTI mechanical engineering for all their hard work in preparing and performing the crash tests included in this study.
Additional information
Funding
Related Research Data
References
- Bauer K, Graw M, Schick S, Willinger R, Peldschus S. 2016. Computer-aided reconstruction of bicycle accidents: determination of head loads and head injury risks. Rechtsmedizin. 26(2):109–114.
- Björnstig U, Näslund K. 1984. Pedal cycling accidents - mechanisms and consequences. A study from Northern Sweden. Acta Chir Scand. 150(5):353–359.
- Björnstig U, Öström M, Eriksson A, Sonntag-Öström E. 1992. Head and face injuries in bicyclists—with special reference to possible effects of helmet use. J Trauma. 33(6):887–893.
- Boele-Vos MJ, Van Duijvenvoorde K, Doumen MJA, Duivenvoorden C, Louwerse WJR, Davidse RJ. 2017. Crashes involving cyclists aged 50 and over in the Netherlands: an in-depth study. Accid Anal Prev. 105:4–10.
- Eppinger R, Sun E, Bandak F, Haffner M, Khaewpong N, Maltese M, Kuppa S, Nguyen T, Takhounts E, Tannous R, et al. 1999. Development of improved injury criteria for the assessment of advanced automotive restraint systems – II. National Highway Traffic Safety Administration.
- Eriksson J, Forsman Å, Niska A, Gustafsson S, Sörensen G. 2019. An analysis of cyclists' speed at combined pedestrian and cycle paths. Traffic Inj Prev. 1–6. DOI: 10.1080/15389588.2019.1658083.
- Fahlstedt M. 2015. Numerical accident reconstructions: a biomechanical tool to understand and prevent head injuries [doctoral thesis, comprehensive summary]. Stockholm: TRITA-STH: report, KTH Royal Institute of Technology;
- Gennarelli TA, Thibault LE, Ommaya AK. 1972. Pathophysiologic responses to rotational and translational accelerations of the head. Paper Presented at: 16th Stapp Car Crash Conference; Detroit, MI.
- Government Offices of Sweden. 2017. A national cycling strategy for more and safer cycling – contributing to a sustainable society and a high quality of life throughout the country. Stockholm: Ministry of Enterprise and Innovation.
- Holbourn A. 1943. Mechanics of head injuries. The Lancet. 242(6267):438–441.
- Margulies SS, Thibault LE. 1992. A proposed tolerance criterion for diffuse axonal injury in man. J Biomech. 25(8):917–923.
- Niska A, Eriksson J. 2013. Cycling accident statistics. Background information to the common policy strategy for safe cycling. Linköping: Swedish National Road and Transport Research Institute.
- Nyberg P, Björnstig U, Bygren LO. 1996. Road characteristics and bicycle accidents. Scand J Public Health. 24(4):293–301.
- Ommaya AK. 1995. Head injury mechanisms and the concept of preventive management: a review and critical synthesis. J Neurotrauma. 12(4):527–546.
- Scheiman S, Moghaddas HS, Björnstig U, Bylund PO, Saveman BI. 2010. Bicycle injury events among older adults in Northern Sweden: a 10-year population based study. Accid Anal Prev. 42(2):758–763.
- Schepers P, Agerholm N, Amoros E, Benington R, Bjørnskau T, Dhondt S, de Geus B, Hagemeister C, Loo BPY, Niska A. 2015. An international review of the frequency of single-bicycle crashes (SBCs) and their relation to bicycle modal share. Inj Prev. 21(E1):e138–e143.
- Stigson H, Fahlstedt M, Svensson MY. 2016, November, 2–4. Preventing Shoulder Injuries in Bicycle Crashes. International Cycling Safety Conference (ICSC); Bologna, Italy.
- Stigson H, Krafft M, Rizzi M, Kullgren A. 2014, November 18–19. Shoulder injuries in single bicycle crashes. International Cycling Safety Conference; Gothenburg, Sweden.
- Swedish Standard Institute. 2012. Helmets for pedal cyclists and for users of skateboards and roller skates. Stockholm, Sweden: Swedish Standard SS-EN 1078:2012.
- Watson JW. 2010. Investigation of cyclist and pedestrian impacts with motor vehicles using experimentation and simulation [PhD thesis]. Cranfield: School of Applied Science, Cranfield University.
- Weijermars W, Bos N, Stipdonk HL. 2016. Serious road injuries in The Netherlands dissected. Traffic Inj Prev.17(1):73–79.
- Zhang L, Yang KH, King AI. 2004. A proposed injury threshold for mild traumatic brain injury. J Biomech Eng. 126(2):226–236.
![Supercharge Your Next Research Paper: Research and write your next paper with Jenni AI.]({"type":"image" , "src" : "/sda/112446/MPU-L1EB.png"})