Reason: Kalman-Bucy phase Block B: radial-truncation proof of global existence under a priori bounds; internally reviewed and validated; batch-approved by Aaron on 2026-07-31.
so, using ∣y∣∣x∣1−∣y∣1=∣x∣∣y∣−∣x∣≤∣x∣d(x,y) and ∣x∣>s,
d(π(x),π(y))≤s∣x∣d(x,y)+s∣x∣d(x,y)≤2d(x,y).
If ∣x∣≤ρ+1<∣y∣: π(y) lies on the segment from 0 to y with ∣π(y)∣=ρ+1, and d(y,π(y))=∣y∣−(ρ+1)≤∣y∣−∣x∣≤d(x,y), so d(π(x),π(y))=d(x,π(y))≤d(x,y)+d(y,π(y))≤2d(x,y); the case ∣y∣≤ρ+1<∣x∣ is symmetric.
Composition: if h:[a,b]→Rk has continuous components, so does π∘h: given a point t and ε>0, choose δ>0 so that ∣s−t∣<δ makes every ∣hl(s)−hl(t)∣<ε/(2k); then for each component i,
The bound ∣x∣≤ρ. The function t↦∣x(t)∣ is continuous (reverse triangle inequality and componentwise continuity, as above). Let
S={c∈[a,b]:∣x(t)∣≤ρ+1for all t∈[a,c]},c∗=supS.
S contains a because ∣x(a)∣=∣ξ∣≤ρ (hypothesis (iii)); S is an interval containing [a,c] for each of its members c; and by continuity ∣x∣≤ρ+1 on [a,c∗], so c∗∈S. On [a,c∗] we have π(x(r))=x(r), so if c∗>a the restriction of x is a solution on [a,c∗] in the sense of the statement, and hypothesis (iii) gives ∣x(t)∣≤ρ on [a,c∗]; if c∗=a this bound holds at t=a anyway. If c∗<b, then since ∣x(c∗)∣≤ρ and ∣x∣ is continuous, there is c′∈(c∗,b] with ∣x∣<ρ+1 on [c∗,c′], whence c′∈S, contradicting c∗=supS. Therefore c∗=b, and by the argument just given (now on [a,b]), ∣x(t)∣≤ρ for all t∈[a,b].
Conclusion. Since ∣x∣≤ρ≤ρ+1 everywhere, π∘x=x, so F(r,x(r))=F(r,x(r)) and x is a solution on [a,b] of the original equation, with ∣x∣≤ρ. Conversely, if h is any solution on [a,b], hypothesis (iii) gives ∣h∣≤ρ, hence π∘h=h and h satisfies the truncated equation; by the uniqueness in Global Existence and Uniqueness for Lipschitz Ordinary Differential Equations in Integral Form, h=x. This proves existence and uniqueness of the solution x=x. ■