6-DOF Interceptor (CX.EV.NW)

ADVANTAGES1 · dim 16

SolvSRK wins. At the comparison noise level, SolvSRK beats the best baseline by at least 10 percentage points of survival, or by at least 0.05 balanced score when survival is tied. Use SolvSRK for this class of problem. All verdicts →

6-DOF Interceptor (CX.EV.NW) benchmark in the defense-autonomy domain.

Defense autonomy

Problem definition

Canonical benchmark implementation

Canonical RHS excerpt from the registered callable used for this benchmark cell. Expand it to verify the state equations; it is not a standalone runnable fixture.

Show canonical RHS excerpt
def rhs(t, y):
    x, yp, z = y[0], y[1], y[2]
    u, v, w = y[3], y[4], y[5]
    phi, theta, psi = y[6], y[7], y[8]
    p, q, r = y[9], y[10], y[11]
    de, da, dr, dc = y[12], y[13], y[14], y[15]

    d = np.empty(16)

    cphi, sphi = np.cos(phi), np.sin(phi)
    cth, sth = np.cos(theta), np.sin(theta)
    cpsi, spsi = np.cos(psi), np.sin(psi)

    V_body = np.sqrt(max(u**2 + v**2 + w**2, 1.0))
    alpha = np.arctan2(w - wind_z, max(abs(u - wind_x), 1.0))
    beta = np.arcsin(np.clip(v / V_body, -0.99, 0.99))
    qbar = 0.5 * _RHO * V_body**2

    CL = CL_alpha * alpha
    CD = CD0 + CD_alpha2 * alpha**2
    L = qbar * S_ref * CL
    D = qbar * S_ref * CD
    Y = qbar * S_ref * 0.5 * beta

    # Body-frame aero forces
    X_aero = -D * np.cos(alpha) + L * np.sin(alpha)
    Y_aero = -Y
    Z_aero = -D * np.sin(alpha) - L * np.cos(alpha)

    # Moments
    L_aero = qbar * S_ref * 0.3 * (Cl_delta_a * da + Cl_p * p * 0.3 / max(V_body, 1.0))
    M_aero = qbar * S_ref * 0.3 * (Cm_alpha * alpha + Cm_delta_e * de + Cm_q * q * 0.3 / max(V_body, 1.0))
    N_aero = qbar * S_ref * 0.3 * (Cn_beta * beta + Cn_delta_r * dr + Cn_r * r * 0.3 / max(V_body, 1.0))

    # Translational dynamics (body frame)
    d[3] = X_aero / m - (q * w - r * v) - _G * sth
    d[4] = Y_aero / m - (r * u - p * w) + _G * cth * sphi
    d[5] = Z_aero / m - (p * v - q * u) + _G * cth * cphi

    # Rotational dynamics
    d[9]  = (L_aero - (Izz - Iyy) * q * r) / Ixx
    d[10] = (M_aero - (Ixx - Izz) * p * r) / Iyy
    d[11] = (N_aero - (Iyy - Ixx) * p * q) / Izz

    # Euler angle kinematics
    sec_th = 1.0 / max(abs(cth), 0.01) * np.sign(cth) if abs(cth) < 0.01 else 1.0 / cth
    d[6] = p + (q * sphi + r * cphi) * sth * sec_th
    d[7] = q * cphi - r * sphi
    d[8] = (q * sphi + r * cphi) * sec_th

    # Position (NED)
    d[0] = cth * cpsi * u + (sphi * sth * cpsi - cphi * spsi) * v + (cphi * sth * cpsi + sphi * spsi) * w
    d[1] = cth * spsi * u + (sphi * sth * spsi + cphi * cpsi) * v + (cphi * sth * spsi - sphi * cpsi) * w
    d[2] = -sth * u + sphi * cth * v + cphi * cth * w

    # Guidance: PN towards target
    tgt_x = tgt_x0 + V_tgt * np.cos(omega_tgt * t) * t
    tgt_y = tgt_y0 + V_tgt * np.sin(omega_tgt * t) * t
    tgt_z = tgt_z0
    dx_t = tgt_x - x
    dy_t = tgt_y - yp
    dz_t = tgt_z - z
    R_los = np.sqrt(dx_t**2 + dy_t**2 + dz_t**2 + 1.0)
    los_el = np.arcsin(np.clip(-dz_t / R_los, -0.99, 0.99))
    los_az = np.arctan2(dy_t, dx_t + 1e-10)

    # LOS error with rate limiting to prevent singularity in crossing geometry
    t_go = max(R_los / max(V_body, 1.0), 0.1)
    los_rate_el = np.clip((los_el - theta), -0.5, 0.5) / t_go
    los_rate_az = np.clip((los_az - psi), -0.5, 0.5) / t_go

    # PN acceleration commands (rate-limited)
    a_cmd_z = N_pn * V_body * los_rate_el
    a_cmd_y = N_pn * V_body * los_rate_az
    de_cmd = np.clip(-a_cmd_z / max(abs(Cm_delta_e * qbar * S_ref * 0.3 / Iyy), 0.1), -0.5, 0.5)
    dr_cmd = np.clip(a_cmd_y / max(abs(Cn_delta_r * qbar * S_ref * 0.3 / Izz), 0.1), -0.5, 0.5)

    # Actuator dynamics (1st-order lag)
    d[12] = (de_cmd - de) / tau_act
    d[13] = (0.0 - da) / tau_act  # wings-level
    d[14] = (dr_cmd - dr) / tau_act
    d[15] = (0.0 - dc) / tau_act

    return d
Parameters
  • CD0 = 0.15
  • CD_alpha2 = 2
  • CL_alpha = 8
  • Cl_delta_a = 1.5
  • Cl_p = -0.5
  • Cm_alpha = -4
  • Cm_delta_e = -6
  • Cm_q = -3
  • Cn_beta = -1
  • Cn_delta_r = 2
  • Cn_r = -0.8
  • Ixx = 0.05
  • Iyy = 0.8
  • Izz = 0.8
  • N_pn = 4
  • S_ref = 0.02
  • V_tgt = 25
  • _G = 9.81
  • _RHO = 1.225
  • m = 5
  • omega_tgt = 0.5
  • tau_act = 0.02
  • tgt_x0 = 877.58256189
  • tgt_y0 = 479.425538604
  • tgt_z0 = -50
  • wind_x = 0
  • wind_z = 0
Initial condition
y(0) = [0, 0, -50, 120, 0, 0, …] [shape=(16,), min=-50, max=120]
Horizon
t ∈ [0, 10]

Canonical RHS excerpt captured from the same registered callable used for the published benchmark. Frozen closure values are summarized below; helper imports and solver settings are intentionally omitted.

Fingerprint

Spread: high

Default noise: high

Recommendation snapshot

Clean best: SolvSRK

Noisy best: SolvSRK

Coverage

14 solver arms · clean + 5 noise levels

Ranked on survival, precision, and speed

Versions & freeze

Methodology →
Freeze
2026-08-13
libsolvsrk
2.3.0
SciPy
1.14
SUNDIALS
CVODE (bundled backend)

20 seeds/cell default · 14 arms · TRL 4–5 · simulation-lab validated · this page: 6-DOF Interceptor (CX.EV.NW) (6-dof-interceptor-cx-ev-nw)

Governed SolvTune benchmark freeze; per-arm medians only. RHS definitions and raw trial rows are not published.

Self-reported by Resonix Labs · not independently verified

Results matrix

Pick an objective and a noise level to rank all arms on survival, median SCD, median nfev, and median wall time. Medians across seeds.

Objective

Best overall trade-off of survival, precision, and speed.

Noise level

#SolverSurvivalSCDnfevWallScore
1SolvSRK
100%
6.8188,4216.19 s0.780
2SciPy RadauSciPy
100%
5.0180,2508.02 s0.738
3SciPy DOP853SciPy
100%
5.096,6503.26 s0.737
4FBDFexternal
100%
4.367,0089.52 s0.723
5SciPy LSODASciPy
100%
4.158,2121.78 s0.717
6Vern7external
100%
4.0105,9829.95 s0.715
7Vern9external
100%
4.0156,49811.78 s0.714
8CVODE Adamsexternal
100%
3.937,8281.33 s0.712
9SciPy RK45SciPy
100%
3.955,5381.79 s0.711
10SciPy BDFSciPy
100%
3.776,3374.60 s0.708
11SciPy RK23SciPy
100%
3.7214,0677.59 s0.708
12CVODE BDFexternal
100%
3.653,7141.91 s0.705
13Tsit5external
100%
3.473,2729.08 s0.699
14TRBDF2external
100%
1.394,76810.57 s0.649

At Clean, best balanced arm is SolvSRK.

Values are medians across seeds, measured by Resonix Labs on Resonix hardware and not independently verified; nfev and wall are on reference lab hardware (indicative). Under injected noise only SolvSRK and the SciPy arms are run. How we measure accuracy → · Verification status →

SolvScout · free

Profile your problem for free

This page shows one published benchmark cell. SolvScout fingerprints your ODE, compares it to the full corpus, and recommends a solver with the same survival / precision / speed ranking you see here — including when a SciPy arm wins.

SolvSRK · 30-day trial

Run the winner on your machine

SolvSRK is the stiffness-adaptive integrator behind the SolvSRK column in these tables. Create an account, activate a machine, and take a 30-day trial — same binary you'd ship after purchase.

Cite this page

Replace the access date. Pin the freeze ID and library versions when comparing against a later export. Cite it as what it is — a self-reported vendor benchmark, not an independently verified result. The note field says so; please keep it.

@misc{resonix_evidence_6_dof_interceptor_cx_ev_nw_2026,
  title        = {Resonix Evidence Portal: 6-DOF Interceptor (CX.EV.NW)},
  author       = {{Resonix Labs (Canada) Inc.}},
  year         = {2026},
  howpublished = {\url{https://resonix.tech/evidence/problems/6-dof-interceptor-cx-ev-nw}},
  note         = {Self-reported vendor benchmark; internally generated by Resonix Labs and not independently verified. Accessed YYYY-MM-DD. Freeze 2026-08-13; libsolvsrk 2.3.0; SciPy 1.14.}
}

Related

TRL 4–5 · simulation-lab validated · 398 problems · 14 solver arms · clean + 5 noise levels

Freeze: 2026-08-13 · scipy 1.14 · libsolvsrk 2.3.0 · Methodology

Self-reported by Resonix Labs · not independently verified · Verification status