Z Advanced Computing, Inc.
Autonomous & Self-Driving
Shortcomings of current Self-Driving technology
After about 20 years and $ 100B investment worldwide:
Self-Driving still stuck at Level 3
(out of 5 levels)
-
Cannot handle complex, unexpected, or rare situations/ events
-
Cannot handle nighttime, rain, or fog
-
Not trustworthy: still requires multiple Remote Human Drivers (Remote Assistance, Supervisor, Operator, Tele-Operator, Tele-Driver, Monitor, or Safety Monitor) for each Robotaxi
-
Limited area of operation (Geo-fencing): inconvenient & not scalable
-
Preloaded/ memorized info (map/ traffic signs) becomes misleading, wrong, and dangerous at road repairs, detours, or after storms & road damages
-
Needs lots of training samples (including simulated samples, which are biased and not accurate)
-
Needs lots of GPU/ CPU & energy (in a huge datacenter facility) for training, which is very expensive
The fundamental problem is…
The current Self-Driving AI is based on Neural Nets.
The solution is…
ZAC Cognitive Explainable-AI (CXAI)
ZAC Tech/Platform addressing Situational Awareness
the most challenging part of the Autonomous Driving
ZAC makes Fully-Autonomous Driving (Level 5) a reality!
ZAC major tech & business advantages for Autonomous Driving
low number of
training samples
higher accuracy,
reliability & reproducibility
much less
computation resources
(e.g., CPU/GPU)
much less
energy or battery
much lower
Carbon footprint
reducing cost
(training, installation, operation)
Explainability
of observed features
better
Situational Awareness
handling unexpected, complex, or rare situations
no need to memorize
street sign locations
no need for
Geo-Fencing
no need for
Remote Human Driver
