Autonomous vehicles (AV) have many applications, not just the transport of people.
We take advantage of the experience acquired to build a vehicle that will transport merchandise in a hydroponic nursery.
That implied new challenges in hardware and software.
One thing is to test algorithms and sensors using a scale prototype and a very different one is to use industrial components and build a prototype with specifications dictated by an end customer.
In recent years GeneXus has researched autonomous vehicles, testing different architectures and algorithms. In this industrial prototype, we incorporate computer vision algorithms through strobe cameras capable of requesting depth, LIDAR sensors / emitters, ultrasound sensors, inertial sensors and imperative algorithms combined with artificial intelligence for people detection.
In this talk we will present the challenges that we find in the construction of a vehicle for an industrial production environment and how we overcome the budgetary constraints through hardware, software and pruning.
Note: This text was automatically translated.