Lots of mention about LIDAR here. As well as IR sensors, etc.
Unless Uber comes out with all the technical specifications that is mounted on the car to detect motion and obstruction on the road, all are mere speculations.
The vehicle in question here would be in the custody of the authorities (impounded), so they should be able to investigate the details on it. Uber should provide as much technical specifications available on the vehicle to aid in the investigation (especially if they are certain that it was more the fault of the pedestrian rather than the self-driving vehicle (SDV)).
SDV companies and advocates should probably pool resources where they could buy a parcel of land, simulate it into a city like settings, and perform edge case tests such as people suddenly crossing, animals, objects, lighting conditions. This will allow for better simulation as well as advancement of technology that would soon realize the possibilities of autonomous driving cars on the road as safe and reliable.
Waymo has 91-acres already[0]. They are selling/donating that land to the "California AutoTech Testing, Development and Production Campus", which will expand to 300 acres and allow others to use it too[1].
It doesn't really matter what the car was actually capable of, so much as what it should have been capable of.
Let's say for argument's sake that the car wasn't equipped with whatever it needed in order to see a person crossing the street in the dark. That doesn't excuse Uber: it just means they shouldn't be testing in the dark on public roads.
Unless Uber comes out with all the technical specifications that is mounted on the car to detect motion and obstruction on the road, all are mere speculations.
The vehicle in question here would be in the custody of the authorities (impounded), so they should be able to investigate the details on it. Uber should provide as much technical specifications available on the vehicle to aid in the investigation (especially if they are certain that it was more the fault of the pedestrian rather than the self-driving vehicle (SDV)).
SDV companies and advocates should probably pool resources where they could buy a parcel of land, simulate it into a city like settings, and perform edge case tests such as people suddenly crossing, animals, objects, lighting conditions. This will allow for better simulation as well as advancement of technology that would soon realize the possibilities of autonomous driving cars on the road as safe and reliable.
Update: Added safe and reliable.