Estou tentando rodar esta rotina no R mas esta dando este erro e tambem ele não esta estimando b2, poderiam me ajudar?
 
lrp.fit<-nls(cv ~ (b0 + b1*x + b2*I(x^2))*(x<=x0p)
 +(b0 + b1*x0p + b2*x0p^2)*(x>x0p),
 start=list(b0=10, b1=7, b2=15, x0p=5),
 trace=F)
 
Erro em nls(cv ~ (b0 + b1 * x + b2 * I(x^2)) * (x <= x0p) + (b0 + b1 *  :
  gradiente singular