-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcz_decomp_prediction.m
71 lines (52 loc) · 1.63 KB
/
cz_decomp_prediction.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
function Xbar = cz_decomp_prediction(X0)
n = 2; % number of states
if size(X0.Z,2)-1 > 3
ng_desired = 3;
nc_desired = 2;
order = (ng_desired-nc_desired)/n;
Xt = reduce(X0,'girard',order,nc_desired);
if ~isempty(Xt)
X0 = Xt;
else
nc_desired = 1;
X0 = reduce(X0,'girard',order,nc_desired);
end
end
c = X0.Z(:,1);
G = X0.Z(:,2:end);
A = X0.A;
b = X0.b;
c1 = c(1,:); c2 = c(2,:);
G1 = G(1,:); G2 = G(2,:);
%% f(eta) and its Bounds
eta = sym('eta',[size(G,2) 1]);
inv_eta = interval(-ones(size(G,2),1),ones(size(G,2),1));
func = {@(eta) 3.*(c1+G1*eta) - (c1+G1*eta).^2/7 - 4.*(c1+G1*eta).*(c2+G2*eta)./(4+(c1+G1*eta));...
@(eta) -2.*(c2+G2*eta) + 3.*(c1+G1*eta).*(c2+G2*eta)./(4+(c1+G1*eta))};
grad{n} = [];
gradfunc{n} = [];
J{n} = [];
fupp{n} = [];
floww{n} = [];
HH{n} = [];
for i = 1:n
grad{i} = jacobian(func{i}(eta),eta);
gradfunc{i} = matlabFunction(grad{i},'Vars',{eta});
J{i} = interval(gradfunc{i}(inv_eta));
[~,~,fupp{i},floww{i},HH{i}] = decomp_signstable_modified(func{i},...
inv_eta.sup,inv_eta.inf,J{i}.sup,J{i}.inf);
end
%% Combinations of H matrices
[~,ind1] = min(fupp{1}-floww{1});
[~,ind2] = min(fupp{2}-floww{2});
H = [HH{1}(ind1,:);HH{2}(ind2,:)];
gd_up = [fupp{1}(ind1,:);fupp{2}(ind2,:)];
gd_low = [floww{1}(ind1,:);floww{2}(ind2,:)];
%% Constrained Zonotopes
cf = 1/2*(gd_low + gd_up);
Gf = [H 1/2*diag(gd_up - gd_low)];
Af = [A zeros(size(A,1),size(Gf,2)-size(A,2))];
bf = b;
cz = conZonotope(cf,Gf,Af,bf);
Xbar = cz;
end