Sundog

Alignment Without Sight

Watch the Sundog Alignment Theorem unfold in real time. Each arc marks a different kind of evidence.

The Sundog Project turns indirect signals into usable control: software that doesn't need perfect information to behave intelligently.

Explore the Framework Three-Body Experiment Balance Workbench Pressure Mines Read the Origin Applications Gallery Public LaTeX Notes View Applications GitHub Repository

Application Rail

Signals Becoming Work

Short field notes from confirmed studies, public prototypes, and planned workbenches.

The Sundog Alignment Theorem

The animation above is not decoration. It is a map of the idea in three registers: theorem as math, theorem as design method, and theorem as empirical claim. The controlled result remains narrow: photometric mirror alignment without target-position access.

Step I · Left Parhelion

A Signal Beside The Source

The first bright spot is not the sun.
It is a displaced signature produced by geometry.
That is the premise.

A sundog is useful here because it makes the core move visible: the thing you can use is not always the thing itself. In the field origin, the laser was blocked by the fastener head. In the experiment, the target position is hidden. Sundog begins when occlusion stops being treated as missing data and becomes a readable signal.

Step II · Right Parhelion

The Frame Of Reference

P = alignment target  ·  L = illumination vector
S(x) = shadow projection  ·  τ(x) = torque
H(x) = ∂S/∂τ  (the halo signature)

The second parhelion gives the scene a frame: source, target, obstruction, projection, and action. As math, Sundog asks whether the projection changes coherently when the agent acts. As design method, it says to instrument that relationship instead of demanding full world state up front.

Step III · Upper Tangent Arc

The Coupling Claim

If action changes the signature coherently,
then the hidden target may be inferable.
H(x) names that coupling.

The upper arc is the dangerous part: it tempts a universal theorem. We keep it conditional. A nonzero relationship between action and signal is not magic; it is a candidate observability channel. The mathematical work is to define that coupling tightly enough that it predicts progress instead of merely sounding profound.

Step IV · Lower Tangent Arc

The Procedure Becomes A Test

Field method: occlude, seat, reacquire.
Lab method: scan, seek, track.
No Cartesian target coordinates.

The lower arc turns the origin story into an empirical task. The controller moves a mirrored end-effector and sees detector intensity plus proprioception, not the target coordinates. Its job is to make the reflected beam land where it should by following the signature produced by its own motion.

Step V · Inner Circle (Iris)

The Closed Loop

Photometric terminal intensity: 0.945.
Oracle terminal intensity: 0.936.
U = 526, p = 0.264.

When the iris closes, the claim becomes measurable. In 30 matched MuJoCo scenes, the photometric controller reached terminal accuracy statistically indistinguishable from the target-aware analytic baseline. The trade was time: indirect feedback converged more slowly. The known failure boundary is tight joint limits.

Step VI · Outer Halo

What The Halo Holds

Math: define the coupling.
Method: act from indirect structure.
Claim: one controlled task currently holds.

The outer halo keeps the registers separate. As math, Sundog is a coupling problem. As design method, it is a way to build from partial, indirect signals. As empirical claim, today, it is the photometric mirror-alignment result. The three-body workbench, EyesOnly, Dungeon Gleaner, and Money Bags are application expressions that motivate the next tests.

Indirect Measurement, Direct Results

Most systems do not reveal their truth directly. You infer them from signatures: shadows, feedback, distortions, response curves.

The Sundog Project is a framework for turning those indirect signals into actionable software control. Where conventional approaches demand complete world state, Sundog asks whether the partial signal, the shadow, the torque, the occlusion, already contains enough structure to act.

In the core experiment, a controller aligns a reflected beam without target coordinates, using only sparse photometric feedback. In workbench and product systems, the same pattern informs guarded three-body control under partial local signals, procedural agents acting under occluded state, verb-field NPC behavior, and softbody motion made interpretable through graph signatures.

Why Indirect Signals Matter

Works Where Direct Inspection Is Hard

Occlusion, expense, or design constraints often make full state access impossible. Sundog operates from partial information.

Agents With Incomplete Knowledge

Agents that know less can feel more alive. Sundog enables coherent behavior from compressed state.

Interpretable Proxy Signals

Instead of raw simulation noise, Sundog transforms physical traces into legible metrics: alignment, torsion, deformation, recovery.

Practical Value

Demonstrated across operating-envelope workbenches, procedural roguelikes, physical simulation, and softbody terrain systems in AI, games, simulation, and tooling.

Working Systems

Photometric Alignment
Research Result

Photometric Mirror Alignment

A controller aligns a reflected beam without target coordinates, matching oracle baseline accuracy in controlled experiments.

Learn More →
Three-Body Workbench
Operating-Envelope Study

Three-Body Dynamics

A guarded accelerometer-proxy TRACK controller improves survival over passive and naive local baselines inside a tested high-velocity near-escape pocket, while low-velocity and equal-mass cells remain explicit failure boundaries.

Learn More →
Balance Workbench
Operating-Envelope Study

Sundog Balance

A shadow-derived cart-pole controller maintains balance inside a mapped lighting and delay envelope, while overhead-light and high-delay cells remain explicit degradation boundaries.

Learn More →
EyesOnly
Instrumented Prototype

EyesOnly / Gone Rogue

Procedural roguelike agents act from compressed perception using stop-conditioned action batches.

Learn More →
Dungeon Gleaner
Product Expression

Dungeon Gleaner

NPCs in a dungeon-crawler town drift between work, social, and errand spots by following the gradient of their own unmet needs. No scripted schedules, no behavior trees, no goal-oriented planner. Personality is per-archetype weighting on a shared vocabulary of verbs.

Learn More →
Money Bags
Instrumented Prototype

Money Bags

Softbody terrain system with graph-based telemetry: torsion, deformation, symmetry, and recovery made legible.

Learn More →

Classic AI vs Sundog AI

Same environment. Same task. Different control logic. Observable difference in behavior.

Classic Approach

  • Requires complete world state
  • Direct target coordinates
  • Immediate convergence
  • Brittle under constraints

Sundog Approach

  • Operates from partial information
  • Indirect signal observation
  • Comparable terminal accuracy
  • Graceful under occlusion

Evidence and Metrics

Terminal Target Intensity

[Bar chart: photometric vs oracle]
p = 0.264

Photometric controller reaches comparable terminal accuracy.

Time-to-Convergence

[Line graph: convergence curves]
188 vs 11.5 steps

The cost of indirect feedback is time, not accuracy.

Stress Test Results

[Failure boundary curve]
Joint limits at 1.0 rad

Known failure at tight joint limits.

Application Metrics

[Multi-domain charts]
ThreeBody | EyesOnly | Gleaner | Money Bags

Utility demonstrated across domains.

Ongoing Research

The Sundog Project is an independent applied research initiative. The defensible scientific claim is narrow: photometric mirror alignment without target-position access in a controlled MuJoCo experiment. The broader applications demonstrate practical utility across procedural systems, simulation, and agent design.

We are continuing to formalize the mathematics, strengthen experimental evidence, and explore new application domains.

What We're Sharing

Open repository, reproducible experiments, comprehensive documentation, and stress test results.

What We're Inviting

Independent replication, collaboration on formalization, application-specific studies, and critical review.

Our Posture

Independent applied research project with experimental software lab as secondary identity.

Explore Sundog

GitHub

View the repository

Documentation

Read the docs

Origin

Trace the physical discovery

Applications

Inspect working systems

Videos

Watch demos

Collaborate

Get involved