Replanning for Evasive Autonomy to Counter Threats

**NEW Q&A ADDED:** Q1 (new): Clarifies that target detection, classification, identification, and terminal-approach guidance are NOT required deliverables—proposers should treat these as inputs from existing Government or third-party capabilities. Focus should be on multi-agent planning capabilities only. **KEY CHANGE:** This significantly narrows the scope and reduces development burden by explicitly excluding sensor-to-engagement pipeline tasks.

# Q&A Changes Summary **New Questions Added (9 total):** - Questions 6-10 added comprehensive guidance on utility function design, modeling frameworks, degraded communications assumptions, simulation/integration artifacts, and performance metrics - Questions 1-5 added clarification on Direct-to-Phase-II maturity requirements, operational scenarios, swarm composition, platform constraints, and threat models **Previously Incomplete Answers Now Completed:** - Q1 A1: Now specifies Indo-Pacific scenario assumptions (GPS-denied, kinetic/non-kinetic threats, swarm survivability) - Q3 A1: Now clarifies TAK integration is **required by end of performance period** (previously blank) **Key Technical Clarifications:** - Utility function should be **operator-adjustable weighted combination** (mission success, survivability, target custody) - Proposers must **recommend framework and justify transition path** rather than use prescribed one - Must handle **GPS denial, comms degradation, latency, message loss, network partitioning** - Solutions should target **heterogeneous Group 1-3 platforms, low-cost/attritable, 10-20 mile line-of-sight comms** - Minimum prior evidence: **simulation results or multi-agent autonomy testing** - Focus scenario: **Maritime Find, Fix, Track, Target, Engage, Assess** (100+ NM offshore)

**Key Changes to Q&A:** - Added 3 new substantive questions about Commander's Intent timing, Direct to Phase II feasibility requirements, and integration expectations with TAK/A2E FANTOM - Q1 and Q3 answers remain incomplete (no substantive response provided) - Q2 answer redirects to official DoD/SBA policy documents rather than providing direct guidance - Reorganized existing Q&As; simulation environment and format questions moved to Q2-Q4 with unchanged answers

Added 1 new Q&A clarifying simulation environment selection: Government approved modeling frameworks are acceptable; simulators must validate multi-agent planner performance against threats with reasonable real-world generalization and test-time visualization capability. No specific preferred environments mandated.

This Q&A clarifies that applicants for Topic SOC26BZ04-DV005 Direct to Phase II submissions must consult the USSOCOM Specific Instructions and Technical Volume (Appendix A) on the portal for detailed formatting requirements, page limits, and any SOCOM-specific proposal guidelines beyond standard DoD SBIR/STTR instructions.

New opportunity: Replanning for Evasive Autonomy to Counter Threats