WEBVTT 1 00:00:09.530 --> 00:00:11.620 Sports related concussions have been in the spotlight 2 00:00:11.620 --> 00:00:14.850 ever since the discovery of Chronic Traumatic Encephalopathy 3 00:00:14.850 --> 00:00:18.280 or CTE, which is a degenerative brain disease believed 4 00:00:18.280 --> 00:00:20.420 to be caused by repetitive concussive 5 00:00:20.420 --> 00:00:22.620 or subconcussive level impacts. 6 00:00:22.620 --> 00:00:25.960 Today, we have evidence that nearly 90% of deceased 7 00:00:25.960 --> 00:00:29.100 former football players had some form of CTE. 8 00:00:29.100 --> 00:00:30.680 So to address this problem, 9 00:00:30.680 --> 00:00:32.650 I'm working with the Obel Sports Technologies 10 00:00:32.650 --> 00:00:34.810 to develop a novel football helmet 11 00:00:34.810 --> 00:00:37.230 that will reduce the likelihood of concussion, 12 00:00:37.230 --> 00:00:39.107 brain damage and ultimately CTE. 13 00:00:40.080 --> 00:00:43.350 Now we view a football helmet as a multi-component system, 14 00:00:43.350 --> 00:00:44.900 where each individual component, 15 00:00:44.900 --> 00:00:46.430 as well as the overall system, 16 00:00:46.430 --> 00:00:48.480 have multiple conflicting objectives. 17 00:00:48.480 --> 00:00:50.930 Such as, to maximize the energy absorption, 18 00:00:50.930 --> 00:00:52.790 while minimizing the weight. 19 00:00:52.790 --> 00:00:54.510 We first tried to design our helmet, 20 00:00:54.510 --> 00:00:57.180 following a much more traditional method of three phases. 21 00:00:57.180 --> 00:00:59.230 Namely, the research and design phase, 22 00:00:59.230 --> 00:01:02.280 the building phase and the testing and analysis phase. 23 00:01:02.280 --> 00:01:04.250 And these three phases are repeated 24 00:01:04.250 --> 00:01:06.370 until you arrive at a optimal solution. 25 00:01:06.370 --> 00:01:08.960 However, we learned that you may be able to optimize 26 00:01:08.960 --> 00:01:12.350 individual components but the engineering uncertainties 27 00:01:12.350 --> 00:01:15.340 will prevent your system from truly being optimal. 28 00:01:15.340 --> 00:01:16.710 Therefore, you cannot simply 29 00:01:16.710 --> 00:01:20.910 add together optimal components and have a optimal result. 30 00:01:20.910 --> 00:01:22.210 This traditional method 31 00:01:22.210 --> 00:01:24.940 is also very time consuming and expensive. 32 00:01:24.940 --> 00:01:27.570 For example, it took our team tens of thousands 33 00:01:27.570 --> 00:01:29.280 of dollars and several years 34 00:01:29.280 --> 00:01:31.950 to go through this cycle just one time. 35 00:01:31.950 --> 00:01:35.140 So for my thesis, I'm developing a better framework 36 00:01:35.140 --> 00:01:37.270 for multi-component systems design 37 00:01:37.270 --> 00:01:40.150 that is less expensive, less time consuming 38 00:01:40.150 --> 00:01:43.180 and can better manage the engineering uncertainties. 39 00:01:43.180 --> 00:01:46.600 First, we established targets for the overall system. 40 00:01:46.600 --> 00:01:49.210 A target for a football helmet may be to totally absorb 41 00:01:49.210 --> 00:01:52.980 the impact energy, then we link the individual components 42 00:01:52.980 --> 00:01:54.560 together using a common metric, 43 00:01:54.560 --> 00:01:56.550 such as kinetic energy. 44 00:01:56.550 --> 00:01:59.500 Then, we begin at the top and sequentially design 45 00:01:59.500 --> 00:02:01.910 each individual components with respect to the other 46 00:02:01.910 --> 00:02:03.640 components, as well as respect 47 00:02:03.640 --> 00:02:05.620 to the overall system targets. 48 00:02:05.620 --> 00:02:07.940 And we manage the engineering uncertainties 49 00:02:07.940 --> 00:02:10.540 by including three common types of error in all 50 00:02:10.540 --> 00:02:12.340 of our design calculations. 51 00:02:12.340 --> 00:02:15.140 Now my framework is also modular, to allow us to quickly 52 00:02:15.140 --> 00:02:19.460 update and reiterate until we arrive at a robust design. 53 00:02:19.460 --> 00:02:23.130 Now you can think of a robust design as a ball on a plateau, 54 00:02:23.130 --> 00:02:26.190 whereas an optimal design would be a ball on a peak. 55 00:02:26.190 --> 00:02:28.890 If your problem were to shift at all, due to errors, 56 00:02:28.890 --> 00:02:31.330 your optimal ball would fall off its peak quickly, 57 00:02:31.330 --> 00:02:32.930 resulting in a loss of performance, 58 00:02:32.930 --> 00:02:36.180 whereas that robust ball would remain on the plateau. 59 00:02:36.180 --> 00:02:38.840 Therefore, we desire robust designs 60 00:02:38.840 --> 00:02:40.070 because they allow our problem 61 00:02:40.070 --> 00:02:43.000 to shift without a substantial loss in performance. 62 00:02:43.000 --> 00:02:44.710 And when it's your brain on the line, 63 00:02:44.710 --> 00:02:46.310 you need your helmet to perform. 64 00:02:47.240 --> 00:02:49.080 Now our framework is also generic, 65 00:02:49.080 --> 00:02:50.750 because football is not the only sport 66 00:02:50.750 --> 00:02:52.260 with a concussion problem. 67 00:02:52.260 --> 00:02:53.810 So we want to be able to update 68 00:02:53.810 --> 00:02:56.030 and design helmets for other sports. 69 00:02:56.030 --> 00:02:58.970 But ultimately, my goal is to allow other engineers 70 00:02:58.970 --> 00:03:00.450 to adopt my framework 71 00:03:00.450 --> 00:03:03.770 to save on time, money and errors also. 72 00:03:03.770 --> 00:03:04.603 Thank you.