Almost two years ago missing safety features in the Uber App contributed to the death of my brother. I am a technical product manager and a data science engineer, I can fix this.
Dec 12th, 2015 was a horrible day for everyone in the Allicock family. Corey my youngest brother died in a car crash when Robert Williams, an Uber driver, ran a red light with Corey in the backseat. At about 1:40am the 73 year old driver would crash into a police cruiser that was already speeding through the intersection. The crash injured everyone else, but it took my brother’s life. I am a technical product manager and a data science engineer. In this article I am going to detail out the ways that machine learning, computer vision and a little bit of app based “common sense” could have saved Corey’s life.
Corey was smart, funny, and arguably the life of the party. Even though he loved to have fun, he didn’t take chances with drinking and driving. That’s why he hailed an Uber car the night of Dec 12th after a company party. After the accident, almost everyone in my family was paralyzed with pain, we felt completely helpless. Was this a freak accident? Why were the vehicles in the intersection at the same time? There were no witnesses so we had to wait for forensics to reconstruct the scene to determine who had ran the light, the Uber driver or the police cruiser. We waited almost a year to find out just who was at fault. I am not a big fan of Uber because of the reports concerning their culture and engineering practices, and I have not spoken to any of their representatives about Corey’s death, but it is my responsibility to offer solutions to this problem to honor Corey’s life and to save the lives of countless others. So, here goes.
Computer Vision and a simple Optical Sensor can detect dangerous driving patterns and alert Uber when drivers reaction time to traffic signals has been compromised. With a simple optical sensor on the dash that is connected to the app, Uber will be able to detect how many times a driver runs a red light or speeds through a yellow one. The optical sensor will be calibrated to detect the colors red, yellow and green, or minimalistically just red and yellow. Machine learning will be able to tell if a drivers reaction time is degrading to a dangerous level and simple alerts will let the driver know if they need to correct their driving patterns or get some rest. If the driver is unresponsive to these alerts, Uber can disable the app. My son, Caleb Davis, a college senior studying Math and Mechanical Engineering at Morehouse College, has pioneered similar technology to detect dangerous levels of gamma radiation in nuclear power plants. We can easily re-engineer the sensor to detect yellow and red lights. We have estimated the cost of such a device to be 34 cents.
A periodic test of reaction time can be built into the app, so older drivers can verify that their reaction time is still within the safe range. The driver of Corey’s Uber was 73 years old. Some initial reports said he was 79. I am not sure of his exact age, but there a few things, including age, that effect driver’s reaction time. Presently Uber doesn’t limit the number of hours a driver can drive. This is a big safety issue. The least Uber can do is build a simple visual test to gauge the driver’s reaction time. If the driver’s reaction time seems slow Uber can disable the app until the driver passes the test or the test can prompt a FaceTime call with the driver from Uber Driver Support. Either way there needs to be some accountability for understanding the mental state of the drivers on the road with people’s lives in their hands.
Machine Learning can be used to identify dangerous intersections and give alerts to the driver. Presently the Waze app alerts drivers when there is a red light camera ahead. The same technology can be used to identify more dangerous intersections by adding historical accident data into the system and giving the driver an audio heads up that a dangerous intersection is ahead. Dangerous intersections are often known by many people in the community who are on the road. People who are new to driving or new to the area might not know. Sometimes these intersections are made more safe by implementing red light camera, speed bumps or other methods. Fixing them usually involves a lot of money and process at the local level. In the meantime simple crowd-sourced reporting can give a heads up that they are approaching a dangerous intersection. I am sure the messaging will have to be thought through carefully so we don’t terrify people. But, I am sure we can think of creative ways to tell people to be extra careful. My wife does that while I am driving all the time ( lol ). We have the data and the technology, let’s use it to help make the roads safer.
Machine Learning and the Accelerometer can be used to detect speeding patterns. This is a simple one. Speed affects reaction time. In Corey’s case if any of the vehicles would have been going just a little slower their reaction time would have been enough to save Corey’s life. Knowing if Uber drivers are chronically speeding is important. Once again, Waze already uses the accelerometer to alert drivers if they are going over the speed limit. By recording the speed patterns of drivers and using machine learning to detect unsafe behavior, Uber can give them alerts or institute disciplinary actions if they speed too much. This will make drivers more wary of their driving habits and save lives.
Many people might think, why are you blaming Uber? If all of these features were implemented Corey still may have died. That is correct, there is no guarantee that any of these features would have saved Corey. Robert Williams may have sped into the intersection irregardless of any of this. I can’t bring Corey back, but I can reduce the likelihood that the next car speeding into an intersection will kill another passenger. And I am willing to do that in loving memory of my brother, so someone else’s brother, sister, husband, wife or family member might be spared.
Disclaimer : I do NOT work for Uber and I am not a party to any lawsuit involving Uber or any other party. These opinions and ideas are my own.
Powered by WPeMatico
Gurupriyan is a Software Engineer and a technology enthusiast, he’s been working on the field for the last 6 years. Currently focusing on mobile app development and IoT.