exp(fb-mix): 轮次7负结论——衰减增益到达 23.4→9.4%、刹停兜底碰撞 1.6→57.8%（绕过安全层）；参数默认关闭，历史最佳仍为 gain1.2 无衰减无兜底（23.4%/1.6%/0.454/973ms）

ml/sim/closed_loop.py

@@ -141,7 +141,8 @@ class TeacherOracle:
 
 @torch.no_grad()
 def rollout(model, kin, tasks, device, K=8, safety=False, seed=0, max_sec=None, return_traj=False, use_sub=True, sub_override=None,
-            qp0=None, steps=1, z_scale=1.0, select='consist', oracle=None, replan='r3', sub_rule='cross', fb_gain=0.0):
+            qp0=None, steps=1, z_scale=1.0, select='consist', oracle=None, replan='r3', sub_rule='cross',
+            fb_gain=0.0, fb_scale=0.0, fb_near=0.2, fb_fallback=0.0):
     """use_sub=False：不用子目标（sf=0）；sub_override=(subs [B,M,6] 绝对关节角, n [B] 个数 或 need [B] bool)：冷启动时用给定途经点代替子目标头；
     qp0 [B,6]：起点的上一帧（默认 = q0，即静止起步）。steps：MeanFlow 采样段数；z_scale：噪声缩放；
     select：consist（部署口径：冷启动价值前瞻、之后与上一条已提交计划的一致性）/ value（每次都价值前瞻）/ first（不选，取第 0 条）。
@@ -191,7 +192,13 @@ def rollout(model, kin, tasks, device, K=8, safety=False, seed=0, max_sec=None,
                 act, val, sub = model.sample_k(feat, K, z, steps=steps)
             inc = act.float().clamp(-1, 1) * astep
             if fb_gain > 0:   # 实验（2026-10-20）：目标反馈加在所有候选上（安全层也评估带反馈的候选）
-                inc = inc + fb_gain * dls_steps(kin, qe, T, R, n=1)[:, :, None, :]
+                fb = dls_steps(kin, qe, T, R, n=1)
+                if fb_scale > 0:   # 距离衰减：远离目标全增益，贴近目标减弱（减少过冲/刹停）
+                    dg, _ = goal_err(kin, qe, T, R)
+                    gdec = fb_gain * torch.clamp(dg / fb_scale, fb_near, 1.0)
+                    inc = inc + gdec[:, None, None, None] * fb[:, :, None, :]
+                else:
+                    inc = inc + fb_gain * fb[:, :, None, :]
             if cold:
                 need = sub[:, 12] > 0.5
                 subs = q[:, None] + sub[:, :12].float().view(B, 2, 6) * 180
@@ -240,6 +247,10 @@ def rollout(model, kin, tasks, device, K=8, safety=False, seed=0, max_sec=None,
                     take = okk & ~ok_any
                     chosen = torch.where(take[:, None, None], ck, chosen); chosen_k = torch.where(take, order[:, kk], chosen_k); ok_any |= okk
                 commit = torch.where(ok_any[:, None, None], chosen, torch.zeros_like(chosen))
+                if fb_gain > 0 and fb_fallback > 0:   # 全拒时的兜底：半速朝目标走一步（须通过安全认证），避免原地停滞
+                    fb1 = dls_steps(kin, qe, T, R, n=1) * fb_fallback
+                    ok1 = kin.gaps(qe + fb1[:, 0], 0.5)['gm'].min(-1).values > 0
+                    commit = torch.where((~ok_any & ok1)[:, None, None], fb1.expand(-1, 3, -1), commit)
                 brakes += (~ok_any).float()
                 pick = chosen_k; committed = ok_any
             if near.any():