1:In the file of dmp.py. in the def step() : self.ddy[d] = (self.az[d] * (self.bz[d] * (self.goal[d] - self.y[d]) - self.dy[d] / tau) + f) * tau in the paper of Ijspeert(2013) the formula is τy¨ = αz(βz(g − y) − y˙) + f , a little different from your formula,i did not understand your formula
2: in the file of dmp_discrete.py why the h of psi is defined as self.h = np.ones(self.n_bfs) * self.n_bfs**1.5 / self.c /self.cs.ax.
3: in the file of dmp.py in the def imitate_path(): f_target[:, d] = (ddy_des[d] - self.ay[d] * (self.by[d] * (self.goal[d]- y_des[d]) - dy_des[d])) without the item of tau.
4:after change the timesteps and runtime of data the result can't follow the original data
1:In the file of dmp.py. in the def step() : self.ddy[d] = (self.az[d] * (self.bz[d] * (self.goal[d] - self.y[d]) - self.dy[d] / tau) + f) * tau in the paper of Ijspeert(2013) the formula is τy¨ = αz(βz(g − y) − y˙) + f , a little different from your formula,i did not understand your formula
2: in the file of dmp_discrete.py why the h of psi is defined as self.h = np.ones(self.n_bfs) * self.n_bfs**1.5 / self.c /self.cs.ax.
3: in the file of dmp.py in the def imitate_path(): f_target[:, d] = (ddy_des[d] - self.ay[d] * (self.by[d] * (self.goal[d]- y_des[d]) - dy_des[d])) without the item of tau.
4:after change the timesteps and runtime of data the result can't follow the original data