There is probably not an analytic solution to your system.
The best way to proceed is to use matlabFunction to create an anonymous function for your system, then use ode45 or one of the others to integrate it.
Your slightly revised code becomes:
syms x1 x2 x3 x4 t u1 u2 real
% syms x1(t) x2(t) x3(t) x4(t) u1 u2 real
x0=[ 0; 0; 0; 0];
% q1=x1(t);
% dq1=x2(t);
% q2=x3(t);
% dq2=x4(t);
q1 = x1;
dq1 = x2;
q2 = x3;
dq2 = x4;
dq = [dq1;dq2];
alfa=6.7;beta=3.4;
epc=3.0;eta=0;
m1=1;l1=1;
lc1=1/2;I1=1/12;
g=9.8;
e1=m1*l1*lc1-I1-m1*l1^2;
e2=g/l1;
H=[alfa+2*epc*cos(q2)+2*eta*sin(q2),beta+epc*cos(q2)+eta*sin(q2);
beta+epc*cos(q2)+eta*sin(q2),beta];
C=[(-2*epc*sin(q2)+2*eta*cos(q2))*dq2,(-epc*sin(q2)+eta*cos(q2))*dq2;
(epc*sin(q2)-eta*cos(q2))*dq1,0];
G=[epc*e2*cos(q1+q2)+eta*e2*sin(q1+q2)+(alfa-beta+e1)*e2*cos(q1);
epc*e2*cos(q1+q2)+eta*e2*sin(q1+q2)];
tol(1)=u1;
tol(2)=u2;
ddq=inv(H)*(tol'-C*dq-G);
% ode1 = diff(x1) == x2(t);
% ode2 = diff(x2) == ddq(1);
% ode3 = diff(x3) == x4(t);
% ode4 = diff(x4) == ddq(2);
ode1 = x2;
ode2 = ddq(1);
ode3 = x4;
ode4 = ddq(2);
odes=[ode1;ode2;ode3;ode4];
ODESYS = matlabFunction(odes, 'Vars',{t [x1 x2 x3 x4 u1 u2]});
Then use ‘ODESYS’ (call it whatever you wish) with the numeric solver. Here, the column vector ‘in2’ contains all the arguments in 'Vars', corresponding to that variable list. See the documentation for matlabFunction to understand what it does.
