2 ms·
Being stubbornly against lidar makes no sense. The sensors are slightly more expensive now, but dropping fast. I want my driving software to have super human ca
by xnx 6y ago
Being stubbornly against lidar makes no sense. The sensors are slightly more expensive now, but dropping fast. I want my driving software to have super human capabilities. Any sensor that can cost effectively save lives and improve performance should be used.
- OnlyOneCannolo 6y agoTesla obviously doesn't see a cost-effective way to incorporate lidar right now. Sometimes you have to make strategic decisions about where to focus your resources.
- dogma1138 6y agoTheir view on lidar is one of the few things I can agree with them regarding autonomous driving. Lidar is essentially a deadend regardless of the cost of the sensors. Lidar’s are good for rudimentary navigation where essentially you use them as high amplitude analogue signals basically is there a wall or a large object in front of me or not. Beyond that lidars are extremely noisy sensors and extremely computationally expensive once you start to use them as a 3D lidar, then you have issues with contested environments, reflectors and many other issues. And most importantly cant provide the majority of information humans use when driving, they can’t see shadows, they can’t see colors, they can’t see lane markings, road signs, car lights traffic lights all which are much more important to provide both human level and super human autonomous driving capabilities. Would sensor fusion improve Tesla right now? Maybe but I’m not 100% it’s the way forward at least not with lidar as a primary sensor, and for emergency collision prevention secondary sensors there are better sensors out there including radars.