Page - 56 - in Proceedings of the OAGM&ARW Joint Workshop - Vision, Automation and Robotics

Fig. 7. Images of the ZED stereo camera showing how light conditions affect the visual cone appearance. circle gets darker and less visible. The Lidar suffers from similar issues, as can be seen in Fig. 8. The cone detection algorithm was tested in an outdoor environment during rain. Since dealing with environmental conditions is part of the challenge of Formula Student, the vehicle will have to run regardless of the current weather. The small axis markers show laser scan points that hit a cone in the vicinity of the robot. The small red dots represent points relative to the robot where an obstacle was detected. The white visual markers on the field show the cones the robot itself is currently detecting; the numbered axis markers show the cones mapped. Usually the field would be empty at the sensor’s maximum range, but during rain, reflections are measured at some point, resulting in a noisy outer cloud structure. Near the robot some of the rays are instantly reflected and reduce the amount of rays that hit a cone, thus making accurate detection more difficult. Fig. 8 shows that the algorithm used is able to map cones with decreasing accuracy but at higher distances. Since scan data has to be compared over time until a cone can be classified correctly, the speed of the detection algorithm is also adversely affected. One can expect the effect of small rocks on the track to have similar effects on detection, as at high speeds they are usually sprayed across the track by the vehicle. VI. CONCLUSION In this paper the TUW Racing team’s approach to the FSD 2017 competition was presented. Details were provided on its hardware and the approach to its control software. The team will begin testing within the next month on a test site with road conditions that resemble the event location and a setup with obstacles corresponding to those detailed in the rules. Initially the team will validate the accuracy of the motion model and perception system in separate road sections. Then it will complete entire tracks at low speeds before gradually increasing speeds while improving controller parameters. While TUW Racing’s main goal this year was successful participation in the first Formula Student Driverless event, its goal for further seasons will extend to improving the efficiency of the vision algorithms and accuracy of the un- derlying motion model. The choice of the sensors, actuators vehicle mapped cones detected cones measurment noise, raindrops new mapped cone Fig. 8. ROS Rviz view of an outdoor test with the pioneer during rain. The laser scan (red) is disturbed by the raindrops. and computers as well as the software framework will be influenced by performance in this year’s competition. ACKNOWLEDGMENT The project is supported by TTTech Automotive GmbH. REFERENCES [1] “How Google’s Self-Driving car works,” Accessed: 14-March- 2017. [Online].Available:http://spectrum.ieee.org/automaton/robotics/ artificial-intelligence/how-google-self-driving-car-works [2] M. Bojarski, D. D. Testa, D. Dworakowski, B. Firner, B. Flepp, P. Goyal, L. D. Jackel, M. Monfort, U. Muller, J. Zhang, X. Zhang, J. Zhao, and K. Zieba, “End to end learning for self- driving cars,” CoRR, vol. abs/1604.07316, 2016. [Online]. Available: http://arxiv.org/abs/1604.07316 [3] Formula Student Germany e.V. Formula Student Rules. Accessed: 14-March-2017. [Online]. Available: https://www.formulastudent.de/ fsg/rules/ [4] T. Howard, M. Pivtoraiko, R. Knepper, and A. Kelly, “Model- Predictive Motion Planning: Several Key Developments for Au- tonomous Mobile Robots,” Robotics Automation Magazine, IEEE, vol. 21, no. 1, pp. 64–73, March 2014. [5] S. Ingle and M. Phute, “Tesla Autopilot : Semi Autonomous Driving, an Uptick for Future Autonomy,” vol. 3, no. 9, Sep 2016. [6] M. Macsek, “Mobile Robotics: EKF-SLAM using Visual Markers for Vehicle Pose Estimation,” Master’s thesis, Vienna University of Technology, 2016. [7] M. Quigley, K. Conley, B. Gerkey, J. Faust, T. Foote, J. Leibs, R. Wheeler, and A. Y. Ng, “ROS: an open-source Robot Operating System,” in ICRA Workshop on Open Source Software, 2009. [8] S. Thrun, W. Burgard, and D. Fox, Probabilistic Robotics (Intelligent Robotics and Autonomous Agents). The MIT Press, 2005. [9] S. Thrun and et al., “Stanley: The robot that won the DARPA Grand Challenge,” Journal of Field Robotics, vol. 23, no. 9, pp. 661–692, 2006. [Online]. Available: http://dx.doi.org/10.1002/rob.20147 [10] G. Todoran and M. Bader, “Expressive navigation and Local Path- Planning of Independent Steering Autonomous Systems,” in 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Oct 2016, pp. 4742–4749. [11] H. Yong and X. Jianru, “Real-time traffic cone detection for au- tonomous vehicle,” in 2015 34th Chinese Control Conference (CCC), July 2015, pp. 3718–3722. [12] S. Zug, C. Steup, J. B. Scholle, C. Berger, O. Landsiedel, F. Schuldt, J. Rieken, R. Matthaei, and T. Form, “Technical evaluation of the Carolo-Cup 2014 - A competition for self-driving miniature cars,” in 2014 IEEE International Symposium on Robotic and Sensors Environments (ROSE) Proceedings, Oct 2014, pp. 100–105. 56