近年來隨著慣性傳感單元這類感測器尺寸變得越來越小，價格也越漸實惠，
開始有公司將他們產品化，開發出安裝在棒球棒底部無線感測器，但其主要的
功能皆為測量球棒的揮棒速度和軌跡（擺動角度）等，並未有偵測球棒打擊位
置的功能。
但在棒球運動中，打者是否具備精準打擊的能力，是影響其打擊表現的關
鍵之一。於是我們想研究出一個擊球位置偵測方法，利用球棒的振動訊號使其
可估算出球棒長軸及圓周方向的撞及位置，為了達成這個目標，我們透過多軸
慣性傳感單元中的加速規收集球棒的振動訊號。
目前在長軸撞擊位置估算方面，在固定撞擊實驗中，以兩種撞擊強度 (1.1J、
1.73J) 收集數據，我們將振動訊號由時域轉成頻域，找出球棒特定的特徵頻率，
並提取出各個頻率的能量峰值作為特徵，使用機器學習的方式進行訓練並估測。
單一撞擊強度下，在 1 公分的容忍誤差內，準確率皆可以接近 75%，運用低強
度 (1.1J) 撞擊建立估算模型，對較高強度 (1.73J、5.70 J) 的撞擊位置進行估算，在 2.5 公分的容忍誤差下，準確率亦可超過 45%，由此可見我們所提出的估算方
法，在一定程度上不受撞擊強度影響。
在圓周撞擊位置預測方面，我們想透過加速規多軸的特性，找出能預測出
球棒不同圓周撞擊位置的方法，目前以 15 度為單位收集五個角度的數據，並利
用不同軸向振動訊號中特徵頻率的峰值間的比值作為特徵進行預測，在單一撞
擊強度下，準確率皆接近 80%。但運用低強度所建立的模型，對較高強度進行
預測，準確率僅有 50%，依然會有難以分辨圓周方向兩側撞擊的問題。

In recent years, with the miniaturization of inertial sensing units and the decreasing price, some companies have started to commercialize wireless sensors installed at the bottom of baseball bats. However, their main functions are to measure the swing speed and trajectory (swing angle) of the bat, and there is no function to detect the hitting position of the bat.
In baseball, whether a batter has the ability to hit accurately is one of the key factors affecting his or her batting performance. Therefore, we want to develop a method for detecting the hitting position of the ball, using the vibration signal of the bat to estimate the collision position of the bat’s long-axis and circumferential direction. To achieve this goal, we collect the vibration signals of the bat through the accelerometer in the multi-axis inertial sensing unit.

Currently, in the prediction of the collision position of the long-axis, we collect data with two impact strengths (1.1J, 1.73J) in a fixed impact experiment. We convert the vibration signal from the time domain to the frequency domain, find the specific eigenfrequency of the bat, and extract the energy peak value of each frequency as a feature. We use machine learning to train and estimate. Under a single impact strength,
the accuracy close to 75% within a tolerance error of 1 cm. We use a low-intensity (1.1J) impact to establish an prediction model and estimate the impact position of higher intensity (1.73J, 5.70 J). Under a tolerance error of 2.5 cm, the accuracy can also exceed 45%. It can be seen that the prediction method we proposed is to some extent not affected by the impact strength.

In the prediction of circumferential impact positions, we aim to utilize the characteristics of multi-axis accelerometers to identify methods for predicting different circumferential impact locations of a baseball bat. Currently, we collect data in increments of 15 degrees and use the peak ratio of characteristic frequencies in vibration signals along
different axes as features for prediction. Under a single impact intensity, the accuracy is consistently close to 80%. However, when applying the model established with low intensity to predict higher intensities, the accuracy drops to only 50%. There still exists a challenge in accurately distinguishing impacts on both sides of the circumferential direction.

[1] H. Brody, “The sweet spot of a baseball bat,” American Journal of Physics, vol. 54,
no. 7, pp. 640–643, 1986.
[2] T. Higuchi, T. Nagami, H. Nakata, M. Watanabe, T. Isaka, and K. Kanosue, “Contribution of visual information about ball trajectory to baseball hitting accuracy,”
PloS one, vol. 11, no. 2, p. e0148498, 2016.
[3] S. Kidokoro, Y. Matsuzaki, and R. Akagi, “Acceptable timing error at ball-bat
impact for different pitches and its implications for baseball skills,” Human Movement Science, vol. 66, pp. 554–563, 2019.
[4] Y. Morishita and T. Jinji, “Accuracy and error trends of commercially available bat
swing sensors in baseball,” Sports, vol. 10, no. 2, p. 21, 2022.
[5] A. T. Bahill, The Science of baseball: Batting, bats, bat-ball collisions, and the
flight of the ball. Springer, 2018.
[6] W.-H. Chen, Y.-C. Feng, M.-C. Yeh, H.-P. Ma, C. Liu, and C.-W. Wu, “Impact
position estimation for baseball batting with a force-irrelevant vibration feature,”
Sensors, vol. 22, no. 4, p. 1553, 2022.
[7] Y. Feng and H. Ma, Detection of Hitting Position of Baseball Bat. National
Tsing Hua University, 2021. [Online]. Available: https://books.google.com.tw/
books?id=y7OgzgEACAAJ
[8] T. Osawa, Y. Tanaka, T. Yanai, and A. Sano, “Position estimation of ball impact
in baseball batting using pvdf films,” in 2017 IEEE World Haptics Conference
(WHC), 2017, pp. 442–447.
[9] R. K. Adair, “Comment on “the sweet spot of a baseball bat,＂by rod cross [am.
j. phys. 66 (9), 772–779 (1998)],” American Journal of Physics, vol. 69, no. 2, pp.
229–230, 2001.
[10] P. Jaramillo, K. S. Manarky, R. S. Adrezin, R. D. Celmer, J. T. Reinard, and
D. Shetty, ““sweet spot＂or “sweet zone＂? modal analysis of a wooden baseball bat for design optimization,” in ASME International Mechanical Engineering
Congress and Exposition, vol. 37130, 2003, pp. 1019–1025.
[11] J. J. Crisco, R. M. Greenwald, J. D. Blume, and L. H. Penna, “Batting performance of wood and metal baseball bats,” Medicine & Science in Sports & Exercise,
vol. 34, no. 10, pp. 1675–1684, 2002.
[12] M. M. Shenoy, L. V. Smith, and J. T. Axtell, “Performance assessment of wood,
metal and composite baseball bats,” Composite Structures, vol. 52, no. 3-4, pp.
397–404, 2001.
[13] R. Quabili, P. Diao, and Y. Mou, “The sweet spot: A wave model of baseball bats,”
UMAPJournal, p. 105, 2010.
[14] G. Vedula and J. Sherwood, “An experimental and finite element study of the relationship amongst the sweet spot, cop and vibration nodes in baseball bats,” The
Engineering of Sport, vol. 5, pp. 626–632, 2004.
[15] H. Brody, “Models of baseball bats,” American Journal of Physics, vol. 58, no. 8,
pp. 756–758, 1990.
[16] L. Van Zandt, “The dynamical theory of the baseball bat,” American Journal of
Physics, vol. 60, no. 2, pp. 172–181, 1992.
[17] A. M. Nathan, “Dynamics of the baseball–bat collision,” American Journal of
Physics, vol. 68, no. 11, pp. 979–990, 2000.
[18] L. Noble and H. Walker, “Baseball bat inertial and vibrational characteristics and
discomfort following ball–bat impacts,” Journal of Applied Biomechanics, vol. 10,
no. 2, pp. 132–144, 1994.
[19] R. K. Adair, “Comment on “how to hit home runs: Optimum baseball bat swing
parameters for maximum range trajectories,＂by gregory s. sawicki, mont hubbard,
and william j. stronge [am. j. phys. 71 (11), 1152–1162 (2003)],” American journal
of physics, vol. 73, no. 2, pp. 184–185, 2005.
[20] J. Albert, J. Bartroff, R. Blandford, D. Brooks, J. Derenski, L. Goldstein, A. P.
Hosoi, G. Lorden, A. Nathan, and L. Smith, “Report of the committee studying
home run rates in major league baseball,” Office of the Commissioner of Baseball,
2018.
[21] R. G. Watts and T. Bahill, “Keep your eye on the ball: curve balls, kunckleballs,
and fallacies of baseball,” (No Title), 2000.
[22] R. Cross and A. M. Nathan, “Scattering of a baseball by a bat,” American Journal
of physics, vol. 74, no. 10, pp. 896–904, 2006.
[23] J. R. Kensrud, A. M. Nathan, and L. V. Smith, “Oblique collisions of baseballs and
softballs with a bat,” American Journal of Physics, vol. 85, no. 7, pp. 503–509,
2017.
[24] A. M. Nathan, “Optimizing the swing,” The Hardball Times, 2015.
[25] ——, “Optimizing the swing, part deux: Paying homage to teddy ballgame.”
[26] D. A. Russell, “Does it matter how tightly you grip the bat?”
[27] STMicroelectronics, “Lsm6dso32 datasheet - stmicroelectronics,” March 2020.
[Online]. Available: https://www.st.com/resource/en/datasheet/lsm6dso32.pdf
[28] R. Pi, “Raspberry pi pico datasheet an rp2040-based microcontroller board,” July
2023. [Online]. Available: https://datasheets.raspberrypi.com/pico/pico-datasheet.
pdf
[29] Tekscan, “Flexiforce™ standard model a301,” 2016. [Online]. Available:
https://www.tekscan.com/sites/default/files/FLX-Datasheet-A301-RevI.pdf
[30] Y. Luo, C. C. Tsang, G. Zhang, Z. Dong, G. Shi, S. Y. Kwok, W. J. Li, P. H. Leong,
and M. Y. Wong, “An attitude compensation technique for a mems motion sensor
based digital writing instrument,” in 2006 1st IEEE International Conference on
Nano/Micro Engineered and Molecular Systems. IEEE, 2006, pp. 909–914.
[31] D. Russel, “Acoustics and vibration of baseball and softball bats,” Acoust Today,
vol. 13, no. 4, pp. 35–42, 2017.
[32] R. Cross, “The sweet spot of a baseball bat,” American Journal of Physics, vol. 66,
no. 9, pp. 772–779, 1998.
[33] L. St, S. Wold et al., “Analysis of variance (anova),” Chemometrics and intelligent
laboratory systems, vol. 6, no. 4, pp. 259–272, 1989.
[34] S. Menard, Applied logistic regression analysis. Sage, 2002, no. 106.
[35] T. Chen and C. Guestrin, “Xgboost: A scalable tree boosting system,” in Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and
data mining, 2016, pp. 785–794.
[36] C.-J. Chen, “Discussion on the recovery coefficient of hard baseball in preventing
injuries from returned balls,” Quarterly of Chinese Physical Education, vol. 15,
no. 3, pp. 69–75, 2001.
[37] V. Bucur, The Acoustics of Wood (1995). CRC press, 2017.
[38] S.-H. Park, W.-Y. Chung, H.-H. Jung, and S.-J. Lee, “Resonance frequency analysis of a baseball bat by impact angle,” Journal of the Korean Wood Science and
Technology, vol. 43, no. 6, pp. 777–783, 2015.