Experimental Robotics Projects Volume Two: Interaction, Embodiment, and Large-Scale Simulation Systems

Experimental Robotics Projects — Volume 2 Interaction, Embodiment, and Large-Scale Simulated Worlds

"Manufacturing is more than just putting parts together. It's coming up with ideas, testing principles and perfecting the engineering, as well as final assembly." — James Dyson

Single-robot thinking breaks the moment systems touch the real world.

A robot operating alone in a controlled environment is a solved problem. What isn't solved — what still collapses in production — is everything that happens when robots interact: with humans, with each other, with objects that push back, at scales where emergent behavior appears without warning.

Volume 1 built the discipline for rigorous simulation-based design. Volume 2 applies it to the hardest class of problems in robotics: interaction under compounded complexity.

Why interaction is where systems fail

The moment a robot makes contact with the world, the problem changes character. Control becomes negotiation. Planning must account for agents that don't follow your model. Learning, which looked stable in isolation, turns brittle when human behavior is in the loop.

Most robotics books treat interaction as an extension of single-robot problems. It isn't. It's a different class of failure — and it requires a different approach to testing.

Volume 2 builds that approach from the ground up.

What this book covers

Every chapter is structured around interaction-driven failure — what causes it, how to detect it, and how to measure it before hardware is involved:

Contact dynamics, friction, compliance, and the trade-offs imposed by morphology

Energy-aware and degradation-aware control over long operational horizons

Human intent modeling and uncertainty-aware interaction under real conditions

Shared autonomy — authority allocation, conflicting commands, and adaptation to operator skill

Trust, transparency, and what happens to both when the robot fails

Learning from demonstration, including what happens when demonstrations are inconsistent

Multi-robot coordination pushed to communication limits and beyond

Emergence, collapse modes, and collective behavior at scale

Large-scale simulation infrastructure for systems that cannot be tested any other way

Nothing is treated as intuitive. Every claim is measured.

Who this is for

Engineers and researchers building systems that must share space, control, or communication with something other than themselves — human-interactive platforms, multi-robot deployments, long-running autonomous systems operating in environments that were never fully specified at design time.

Volume 1 is the prerequisite. Not because the material demands it, but because the discipline does. Volume 2 assumes you've already internalized rigorous simulation-based thinking. It then shows you what that thinking reveals when the problem gets genuinely hard.

If Volume 1 teaches discipline, Volume 2 teaches consequences.

Not theory. Not demos. The measured behavior of systems under conditions they weren't designed for — surfaced in simulation, before deployment makes the cost real.

If your robots must interact, coordinate, or persist at scale, Experimental Robotics Projects — Volume 2 belongs alongside Volume 1.