理想的人機系統操作應該同時考慮到人為操作跟機器的可靠度。根據大部分的資料比較出相較於在機器端的設計改變，人為的操作誤差對於人機系統的影響的參考資料較少。不同的個體使用同一個流程或步驟操作同一台機器或是系統時，得到的結果往往不進相同；即使由同一個個體使用同一個流程操作同一個機器或是系統時，也不一定回得到完全相同的結果。目前大部分的文獻都只是提供事前建議來確保操作者可以達到操作的準確性，而不是提供一個特定的方法來確保操作者在操作的同時可以達到操作的準確性。在本論文中，我們提出了一個補償方法來補償人為的操作不確定因素。系統必須先經過訓練的流程，這時候系統會提供操作者指令和單純收集操作者的操作結果，接下來系統會比較指令和操作結果中間的誤差，建立補償過的指令並傳送給使用者確保使用者可以達到正確的操作，並且用來改善人機操作系統整體的可靠度。在本論文中，我們將這套方法應用在智慧停車輔助系統上面，現存的停車系統在都會區使用時會遇到若干的限制，我們希望可以利用這個新的方法來建構即時的補償系統，用以補償駕駛在操作上的不足。系統會提供駕駛人語音指令引導，並且套用在各種不同情況的停車格中，來驗證論文所提出的方法是否有效。論文中使用的方法即使在駕駛有百分之三十的操作誤差時，都還是可以順利地引導駕駛人停入停車格中，此外我們也有效的降低了停車輔助系統的限制到車長的1.07倍，由此可以得知，補償方法不但可以確保使用者達到正確的操作，也可以提高操作的可靠度。

Successful executions of man-machine systems require consistent human operations and reliable machine performances. Compared with the abundant resources on machine reliability improvements, human-related operational uncertainty that has direct impacts on man-machine systems received little attention. Different individuals performing the same task on the same machine could have different results; even with a single individual, performing the same task multiple times might not guarantee an identical outcome. Most studies and formal documentations only provide suggestions to alleviate human uncertainty instead of providing specific methods to ensure operation accuracy in real time. In this paper we present a method to compensate for human operating uncertainty. This system first learns the response of an operator by providing instructions and collecting outcomes. The operator's behavior pattern can then be constructed and a new compensated instruction will be provided to ensure the operator's final outcome in real time, hereby improve the reliability of the entire man-machine system operation. The effectiveness of the proposed method is demonstrated using an intelligent vehicle parking assist. Existing parking assist systems have various limitations in urban parking due to possible operational uncertainty and system limitations. Our new development computes a theoretical path once a parking space is identified. Audio commands are then sent to the driver with real-time compensation for a minimal deviations from the path. Various real-world urban parking scenarios are used to demonstrate the effectiveness of the proposed method. Our proposed approach can tolerate drivers with up to 30% deviations from the theoretical path. We are also able to park a vehicle with a space that is as small as 1.07 times the vehicle length. Compensated results not only allow more diverse operators to achieve a desired goal, but also ensure a high reliability of meeting such goals.

[1] Y. Wang and M. Cartmell, Autonomous vehicle parallel parking design using function
tting approaches," Robotica, vol. 16, no. 02, pp. 159{170, 1998.
[2] S. Choi, C. Boussard, and B. D'Andrea-Novel, Easy path planning and robust control for
automatic parallel parking," in IFAC World Congress, vol. 18, pp. 656{661, 2011.
[3] Automatic Parking Guidance System (APGS) by ARTC." http://www.artc.org.tw/
english/02_research/02_01detail.aspx?pdid=11
Accessed: 2014-01-21.
[4] Advanced Parking Guidance System by ARTC." http://www.artc.org.tw/chinese/
02_research/02_01detail.aspx?pdid=13
Accessed: 2014-05-12.
[5] Parking Guidance System by ARTC." http://www.artc.org.tw/chinese/02_
research/02_01detail.aspx?pdid=12
Accessed: 2014-05-12.
[6] M-Benz B200 Turbo." http://fayqfs.pixnet.net/blog/post/
25928340-m-benz-b200-turbo-
Accessed: 2014-05-12.
[7] BMW Connected Drive." http://www.bmw.com.tw/com/tw/newvehicles/3series/
touring/2012/showroom/convenience/parking_assistant.html#t=l
Accessed: 2014-05-12.
[8] Ford;Õ^Ê K’êÕÊûq-‡ÕK‹ by Joseph Tao." http://chinese.
engadget.com/2013/03/06/ford-active-system-chinese-hands-on/
Accessed: 2014-04-16.
[9] A. Pentland and A. Liu, Modeling and prediction of human behavior," Neural Computa-
tion, vol. 11, no. 1, pp. 229{242, 1999.
[10] M. Nechyba and Y. Xu, Human control strategy: abstraction, veri cation, and replication,"
Control Systems, IEEE, vol. 17, no. 5, pp. 48{61, 1997.
45
[11] S. Sekizawa, S. Inagaki, T. Suzuki, S. Hayakawa, N. Tsuchida, T. Tsuda, and H. Fujinami,
Modeling and recognition of driving behavior based on stochastic switched arx model,"
Intelligent Transportation Systems, IEEE Transactions, vol. 8, no. 4, pp. 593{606, 2007.
[12] T. Kumagai, Y. Sakaguchi, M. Okuwa, and M. Akamatsu, Prediction of driving behavior
through probabilistic inference," in Proceedings of the Eighth International Conference on
Engineering Applications of Neural Networks, pp. 8{10, 2003.
[13] D. Mitrovic, Reliable method for driving events recognition," Intelligent Transportation
Systems, IEEE Transactions, vol. 6, no. 2, pp. 198{205, 2005.
[14] J.-H. Cheng and Y.-H. Chang, Application of a fuzzy knowledge base on bus operations
under uncertainty," in Fuzzy Systems Conference Proceedings, 1999. FUZZ-IEEE'99. 1999
IEEE International, vol. 3, pp. 1355{1360, IEEE, 1999.
[15] N. Stanton and S. Stevenage, Learning to predict human error: issues of acceptability,
reliability and validity," Ergonomics, vol. 41, no. 11, pp. 1737{1756, 1998.
[16] S. Nigam and J. Turner, Review of statistical approaches to tolerance analysis,"
Computer-Aided Design, vol. 27, no. 1, pp. 6{15, 1995.
[17] T.-C. Hung and K.-Y. Chan, Multi-objective design and tolerance allocation for singleand
multi-level systems," Journal of Intelligent Manufacturing, vol. 24, no. 3, pp. 559{573,
2013.
[18] J. Reason, Modelling the basic error tendencies of human operators," Reliability Engi-
neering & System Safety, vol. 22, no. 1, pp. 137{153, 1988.
[19] M. Lipow, Prediction of software failures," Journal of Systems and Software, vol. 1,
pp. 71{75, 1980.
[20] L.-H. Chen, Wind farm optimization with turbine blade design considering geographical
constraints," Master's thesis, National Cheng Kung University, 2012.
[21] R. Pintelon, P. Guillaume, and J. Schoukens, Uncertainty calculation in (operational)
modal analysis," Mechanical Systems and Signal Processing, vol. 21, no. 6, pp. 2359{2373,
2007.
46
[22] G. Ribas, A. Leiras, and S. Hamacher, Operational planning of oil re neries under uncertainty,"
IMA Journal of Management Mathematics : Special Issue on Applied Stochastic
Optimization, vol. 23, no. 4, pp. 397{412, 2012.
[23] C. Johnson, Why human error modeling has failed to help systems development," Inter-
acting with Computers, vol. 11, no. 5, pp. 517{524, 1999.
[24] W. Rouse and S. Rouse, Analysis and classi cation of human error," IEEE Transactions
on Systems, Man and Cybernetics, vol. SMC-13, no. 4, pp. 539{549, 1983.
[25] J. Kim, Y. Kim, and D. Hwang, Modeling of human driving behavior based on piecewise
linear model," AUTOMATIKA-ZAGREB, vol. 46, no. 1/2, p. 29, 2005.
[26] R. Maxion and R. Reeder, Improving user-interface dependability through mitigation of
human error," International Journal of Human-Computer Studies, vol. 63, no. 1, pp. 25{50,
2005.
[27]
[28] D. Gunia, M. Kalabis, J. Meise, K. C. Muller, N. Rolfes, S. Toledo, and T. Wey, Vehicle
park assist system and method for parking a vehicle using such system," May 1 2012. US
Patent 8,169,341.
[29] Ford Active Park Assist Diagram." http://www.machine-history.com/node/809?
size=_original
Accessed: 2014-04-17.
[30]"停車導引系統及其引導方法." http://twpat-simple.tipo.gov.tw/tipotwoc/
tipotwkm?@@1026905039
Accessed: 2014-04-17.