it’s more than one year since I started this series “where are you in the next 5 years”, I wound love to transfer to “the next 4 years”, and thanks for the opportunity to get back in China, so there is a high chance to involve in the market heavily in a short time.
at the begining of the year, travelled around the whole nation, stay in Shanghai, Beijing, Shenzhen, Guangzhong. and that was a great chance to get familiar with the startups in autonomus vehicle, as this trip really gave me some input, and till now I had another half year in one of the top OEMs in China. combined with this two sources, which gave me the kind of the whole picture of ADS market happening in China. I would love to write this blog more in bussiness thought, rather than engineering way.
L4 startups
ADS leap time is about 2016 to the first half year of 2018. there are a bunch of startups and also most OEMs have build their ADS teams.
the startups, e.g. Pony.ai, WeRide, AutoX, roadstar(the new split), ToSimple, Momenta. which are still very active recently.
Today I have taken PlusAI’s tech open day, I have to say, most of these startups have very similar tech roadmap. I personally, think that’s a really sad thing.
a few teches they all have:
simulation pipeline
data pipeline(collect, label, training)
AI based perception, motion planning
friendly HMI
WeRide and Pony.ai are in Robtaxi service; ToSimple and PlusAI are in highway logistics; Momenta is in harbor transformation.
Alibaba, jingdong, Meituan e.t.c are in small personal package delivery shuttles, similar as Nuro.
OEMs focus in the passenger vehicles.
DiDi focus in taxi services as well, similar as Uber, Waymo.
all of them can be called as ADS service suppliers. however, most of them use the exactly same sensor packages, including Lidar, Camera, Radar, GPS e.t.c. the software stacks during prodcut dev as mentioned above are mostly similar; there maybe a few special features during the services in deployment, e.g. Robtaxi may have a Uber-like call-taxi app e.t.c, Rather than that, nothing really is amazing about ADS itself.
and mostly this is not a tech problem, it’s must be defined or find out by the social guys, who are from the real needs.
in the engineering work environment, it’s easily to misunderstand the role of engineering. engineering is the bumper, only when the house need fixed, there is a need for bumper. However, in an engineering-centered env, it’s so easy to tell no difference between I have the bumper and I have the needs. My experieince till now, I am learning how to use the bumper well, but few thinks why need to learn to use the bumper.
on the other hand, what kind of tech is really helpful or profitable?
by chance, to talk with Unity China team, who are enhancing Unity3D engine with cloud support, unity simualtion, which is the feature I am looking for a while. if the tech pipeline is the waterflow, the Unity team is the one standing at the upper flow, who can implement new features in engine.
just like Nvidia, Google e.t.c, these are the guys who really make a difference with their tech. and it’s profitable of course.