Onboard Autonomy Instead of Central Control
Scott Sanders, Chief Growth Officer at Forterra, on autonomy for every vehicle
This points to a core design rule for military autonomy, the useful system is the one that keeps working after the network breaks. Forterra is building vehicles that carry their own sensing, planning, and compute, because a convoy, robot truck, or battlefield support vehicle cannot wait for a remote server or command hub to tell it what to do. That same edge first design also shows up in yards and terminals, where wireless coverage is uneven and infrastructure is expensive to install.
-
Forterra describes an autonomous ground vehicle as a full computer system on the vehicle itself, with sensor fusion, positioning, communications, and maintenance logic running at the edge. That matters because ground systems have to move through GPS denied, degraded environments while carrying payloads, people, or supplies.
-
The practical alternative is not one central controller for every drone, truck, and ship. It is many autonomous nodes that can still share tasks when links are available. That is also where platforms like Anduril’s Lattice fit best, as a coordination layer for teams of assets rather than a permanent dependency for every action.
-
The commercial parallel is Forterra’s Kalmar partnership on terminal tractors. Ports and distribution yards are messy physical environments with weak networks, changing layouts, and constant vehicle movement, so the autonomy stack has to make local decisions on the machine instead of relying on fixed site infrastructure.
The next step is a battlefield made of semi independent machines, where each vehicle can drive, sense, and survive on its own, but still plug into wider command networks when available. That favors companies that package onboard autonomy with interoperable coordination software, and it pushes the market away from cloud dependent robotics toward rugged edge systems built for contested operations.