[Paper Review] Motion camouflage in three dimensions

Puduru Viswanadha Reddy, Eric W. Justh+1 more|ArXiv.org|Mar 8, 2006

Guidance and Control Systems7 references4 citations

TL;DR

This paper formulates a three-dimensional motion camouflage feedback law using natural Frenet frames and curvature controls, enabling a pursuer to maintain constant bearing relative to a moving evader—biologically plausible and analogous to pure proportional navigation in missile guidance. The key contribution is a high-gain, range-dependent control law that ensures the pursuer appears motion-camouflaged from the evader's perspective, validated through simulations with varying evader trajectories.

ABSTRACT

We formulate and analyze a three-dimensional model of motion camouflage, a stealth strategy observed in nature. A high-gain feedback law for motion camouflage is formulated in which the pursuer and evader trajectories are described using natural Frenet frames (or relatively parallel adapted frames), and the corresponding natural curvatures serve as controls. The biological plausibility of the feedback law is discussed, as is its connection to missile guidance. Simulations illustrating motion camouflage are also presented. This paper builds on recent work on motion camouflage in the planar setting [1]. [1] E.W. Justh and P.S. Krishnaprasad (2005). "Steering laws for motion camouflage," submitted for publication, (see also arXiv:math.OC/0508023v1).

Motivation & Objective

• To develop a biologically plausible, feedback-based control law for motion camouflage in three-dimensional space.
• To extend prior planar motion camouflage work to 3D using natural Frenet frames and curvature-based control.
• To establish a connection between motion camouflage and pure proportional navigation guidance (PPNG) in 3D.
• To simulate and validate the control law under diverse evader motion profiles, including sinusoidal, random, and circular trajectories.
• To test the hypothesis that echolocating bats use a similar strategy during prey capture, based on experimental trajectory data.

Proposed method

• Models pursuer and evader as point particles with position, velocity, and curvature-controlled trajectories using natural Frenet frames.
• Defines the pursuer's dynamics using unit tangent vector $\mathbf{x}_p$, normal vectors $\mathbf{y}_p$, $\mathbf{z}_p$, and curvature controls $u_p$, $v_p$.
• Derives a high-gain feedback law based on the relative bearing angle $\Gamma$, driving it toward $-1$ to enforce constant bearing.
• Uses the relative position vector $\mathbf{r} = \mathbf{r}_p - \mathbf{r}_e$ and its normalized form to compute the control input via $A_M^{MCPG} = \mu \nu_p^2 \left[ \mathbf{x}_p \times (\dot{\mathbf{r}} \times \mathbf{r}/|\mathbf{r}|) \right]$.
• Establishes equivalence to pure proportional navigation by identifying $\Omega_L$ as the line-of-sight angular rate and $V_M$ as the pursuer speed.
• Introduces a dimensionless gain $N^{MCPG} = \mu \nu_p r_o$ to scale the control law, enabling range-dependent gain modulation.

Experimental results

Research questions

• RQ1 How can motion camouflage be systematically formulated and controlled in three-dimensional space using biologically plausible feedback mechanisms?
• RQ2 What is the mathematical relationship between motion camouflage and pure proportional navigation guidance in 3D?
• RQ3 How does the proposed feedback law maintain constant bearing under varying evader trajectories, including non-uniform motion?
• RQ4 What role does range-dependent gain play in stabilizing the motion camouflage behavior, and how does it relate to echolocation in bats?
• RQ5 Can the motion camouflage strategy be extended to finite-point camouflage, where a fixed object is used as a camouflage reference?

Key findings

• The motion camouflage proportional guidance (MCPG) law successfully maintains the relative bearing angle $\Gamma$ near $-1$, ensuring the pursuer appears motion-camouflaged from the evader’s perspective.
• Simulations confirm that the control law stabilizes the system even with sinusoidal or randomly varying evader curvature inputs, producing nearly parallel pursuer-evader baselines.
• For circular evader trajectories, the pursuer converges to a stable, motion-camouflaged path with minimal transient deviation.
• The control law exhibits high gain at long ranges and reduces gain at close range, mimicking the behavior of echolocating bats with known ranging capabilities.
• The MCPG law is mathematically equivalent to pure proportional navigation guidance (PPNG) when scaled by a range-dependent gain factor, establishing a direct link between biological stealth and missile guidance.