Solving a function by interpolation

I'm trying to find the solution x of f(a, b, c, d, x) = 0 where f is a function I implemented. I managed to create an array X = xmin:xstep:xmax; Then I tried to interpolate in order to find x such that f = 0: x = interp1(f(a,b,c,d,X),X,0);

But I sometimes get an error by using interp1 (which I don't get by using a home made slower interpolation function) :

Error using griddedInterpolant The grid vectors must contain unique points.

Error in interp1 (line 151) F = griddedInterpolant(X,V,method);

Any idea how to solve this? If you have any suggestion to get the value x to solve f(a,b,c,d,x) =0 I'm listening too.

Thanks in advance

[KSSV]

Show the line code which gives error.

[Maxandre Coulon]

In my console, I have:

Error using griddedInterpolant

The grid vectors must contain unique points.

Error in interp1 (line 151)

F = griddedInterpolant(X,V,method);

Error in Envelope_30_06 (line 165)

epsilon_teta_0(indexDeltaTau+1,indexDeltaf+1) = interp1(ATAN(epsilon_tau_0_prev(indexDeltaTau+1,indexDeltaf+1), A0, t_centered, corr_vec, A1, DeltaTau(indexDeltaTau+1), 0,

epsilon_teta, epsilon_f_0(indexDeltaTau+1,indexDeltaf+1), Ti, Deltaf(indexDeltaf+1)),epsilon_teta,0);

my ATAN function is :

function ATAN = ATAN(epsilon_tau, A0, t_centered, corr_vec, A1, delta_tau, delta_teta, epsilon_teta, epsilon_f, Ti, Delta_f)
corr_LOS_P = interp1(t_centered,corr_vec,epsilon_tau);
corr_NLOS_P = interp1(t_centered,corr_vec,epsilon_tau + delta_tau);
%IP = A0 / 2 .* sinc(pi .* epsilon_f .* Ti) .* corr_LOS_P .* cos(epsilon_teta + pi .* Ti .* epsilon_f) + A1 / 2 .* sinc(pi .* (epsilon_f + Delta_f) .* Ti) .* corr_NLOS_P .* cos(epsilon_teta + delta_teta + pi .* (epsilon_f + Delta_f) .* Ti);
%QP = A0 / 2 .* sinc(pi .* epsilon_f .* Ti) .* corr_LOS_P .* sin(epsilon_teta + pi .* Ti .* epsilon_f) + A1 / 2 .* sinc(pi .* (epsilon_f + Delta_f) .* Ti) .* corr_NLOS_P .* sin(epsilon_teta + delta_teta + pi .* (epsilon_f + Delta_f) .* Ti);
IP = A0 / 2 .* corr_LOS_P .* cos(epsilon_teta + pi .* Ti .* epsilon_f) + A1 / 2 .* corr_NLOS_P .* cos(epsilon_teta + delta_teta + pi .* (epsilon_f + Delta_f) .* Ti);
QP = A0 / 2 .* corr_LOS_P .* sin(epsilon_teta + pi .* Ti .* epsilon_f) + A1 / 2 .* corr_NLOS_P .* sin(epsilon_teta + delta_teta + pi .* (epsilon_f + Delta_f) .* Ti);
ATAN = atan(QP ./ IP);
end

in interp1, the problem is in

Xqcol = Xq(:);
num_vals = size(V,2);
% some lines
else % can use ND
    if (num_vals > 1)
        Xext = {cast(X,'double'),(1:num_vals)'};
        F = griddedInterpolant(Xext,V,method);
        VqLite = F({cast(Xqcol,class(Xext{1})),Xext{2:end}});
    else
        F = griddedInterpolant(X,V,method);
        VqLite = F(Xqcol);
    end
end

[Sargondjani]

Well the error says it: "The grid vectors must contain unique points." If a particular value of x has two 'y' values, then you can not use interp1 in that way. If x is a function of y, then swap your x and y variables in interp1 and it should work.