"A lot of folks in the industry refer to waypoint following or coordinated pre-programmed flight as autonomy," Ben Wolff, president and CEO of Palladyne AI, told Breaking Defense.
Ben Wolff on autonomy: unmanned is not the same as autonomous
Wolff draws a strict line between “unmanned” systems and genuinely autonomous machines. He argues that many systems labeled autonomous today are merely executing pre-programmed scripts or waypoints — human decisions made in advance — and that true autonomy requires real-time decision‑making by the machine. To make that distinction concrete, Wolff maps drone autonomy to the five levels commonly used for self‑driving cars, from Level 0 (every movement directed by a human) to a Level 5 analogue where a soldier launches a platform, assigns a high‑level mission, and the drone completes it without further human direction.
Even so, Wolff insists autonomy should preserve human authority over critical acts of violence: “The ability to abort or to proceed with an attack should, in my view, always reside with a human,” he said, while acknowledging “our adversaries might not always agree.”
DECA: decentralizing intelligence on the edge
Palladyne AI describes the enabling technology as decentralized embodied collaborative AI, or DECA. In DECA‑equipped drones the intelligence lives on each aircraft rather than in a central controller, enabling each drone to sense, reason and act in the physical world in real time. The company emphasizes closed‑loop sense‑reason‑act systems on the edge so drones can react without waiting for operator instructions.
Wolff uses the everyday analogy of catching a baseball: you don’t consciously calculate arm trajectory, you simply react. He says the same reflexive, split‑second behavior is what DECA brings to drones operating under the unpredictable conditions of a battlefield.
Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scramblePalladyne AI’s SwarmOS and heterogeneous collaboration
To turn independent edge intelligence into a coordinated force, Palladyne AI developed SwarmOS. The software curates and shares high‑value information among disparate platforms so that a heterogeneous swarm — drones from different vendors and with different sensors — can collaborate. Wolff noted the software places algorithms on individual drones, removing the need for centralized control.
SwarmOS also creates a mesh network through which platforms share observations and build a shared picture. In the company’s account of the Army PC‑C6 (Project Convergence Capstone 6) exercise, drones from multiple manufacturers carried the SwarmOS software and collaborated to complete missions while operating with limited sensor information on any single airframe.
Oracle‑Class Wolf Pack Swarming: prediction as a force multiplier
Palladyne brands its predictive approach “Oracle‑Class Wolf Pack Swarming.” Wolff describes the algorithms as predictive, leveraging the sum of knowledge across a swarm to forecast likely target movements or teammate behavior. He offers a concrete example: a swarm tracking a convoy that disappears under a tree canopy can use collective data to estimate where the convoy is likely to reemerge.
That predictive layer is presented as a “holy grail” capability — extending situational awareness from passive observation to anticipatory action that could give operators a strategic edge.
Pilots, the Department of War, and the problem of scale
Wolff frames a practical constraint for widespread swarm deployment: people. He notes demand signals from the Department of War focused on increasing the volume of drones, and questions whether there are enough pilots trained for many different platforms to operate large fleets. Launching a thousand drones, he argues, is not feasible if each requires a dedicated pilot.
SwarmOS’ response is to reduce cognitive load on operators so a small number of pilots can manage large fleets from a single controller, focusing on higher‑level decisions—defining mission parameters, confirming targets, and authorizing or aborting attacks—while the swarm conducts low‑level navigation and sensing autonomously.
What this means for pilots, the Department of War, and commanders
- Pilots: Expect to shift from manual flight control toward supervisory roles — approving engagements, redirecting missions, and relying on a curated, shared picture built by the swarm.
- The Department of War: Will need procurement and training strategies that accept heterogeneous fleets and prioritize software that enables interoperability and reduced operator burden.
- Commanders and warfighters: Stand to gain broader, predictive situational awareness if DECA and SwarmOS perform as described, but will still face the decision of when to retain human authorization for kinetic effects.
Taken together, the arguments from Wolff and Palladyne AI paint a near‑term vision where autonomy is judged not by the absence of a human pilot but by how little cognitive overhead the machines impose on those pilots. The company’s emphasis on edge intelligence, heterogeneous interoperability and predictive algorithms outlines a roadmap from today’s waypoint scripts to a future of collaborative, anticipatory swarms — all while insisting that strategic choices about lethal force remain a human responsibility.
Source: Breaking Defense — Bringing real autonomy to battlefield drone swarms
