Robot Design by Animation

overview

intro

data

model

optimization

simulation

hybrid_1

hybrid_2

conclusions

Simulation

Simulation Results

Figure 13
Stable Trajectory of lo_hipK solution
Figure 14
Stable Trajectory of hi_hipK solution
The following results compare two solutions for X:

simulation results for lo_hipK after 100 iterations (Figure 13)
simulation results for hi_hipK after 600 iterations (Figure 14)

For both results, the model was simulated and allowed to run until a stable trajectory was achieved. This stable cycle was used to compare with the original kinematic data. Finally, a disturbance was added to the simulation (results below).

Overlay Movies

Figure 15
Overlay for lo_hipK (movie)
Figure 16
Overlay for hi_hipK (movie)
Figures 15 and 16 show the kinematic data and the simulation results for both optimization solutions.
As can be seen, both trajectories are qualitatively similar. The trajectory for hi_hipK appears to be out of phase.

Trajectory Comparison

Figure 17
X position for simulated trajectory of lo_hipK
Figure 18
Y position for simulated trajectory of lo_hipK

Figure 19
Orientation for simulated trajectory of lo_hipK

Figure 20
X position for simulated trajectory of hi_hipK

Figure 21
Y position for simulated trajectory of hi_hipK

Figure 22
Orientation for simulated trajectory of hi_hipK

These figures compare the kinematic data with one stride period of the simulations (taken after the simulation has converged to a stable trajectory).
As seen in the figures, the results for lo_hipK appear to match the position trajectories slightly better than the results for hi_hipK. In particular, note that the simulation for hi_hipK in the X direction is out of phase with the kinematic data.

Note that the kinematic data is not regular (non-cyclic), while the simulation results return to the same state at the end of the stride period.

Disturbance Rejection Movies

Figure 23
Disturbance rejection for lo_hipK
Figure 24
Disturbance rejection for hi_hipK
For both solutions, the model was simulated and a large disturbance was introduced during the simulation. The disturbance consisted of the addition of a velocity in the x direction of 1m/s.
Figures 23 and 24 show the resulting motion.

As seen from the movies, the solution for hi_hipK is more robust to the disturbance. The solution for lo_hipK takes about three of four stride cycles to converge back to the particular direction.

Simulation
	Simulation Results
Figure 13 Stable Trajectory of lo_hipK solution Figure 14 Stable Trajectory of hi_hipK solution	The following results compare two solutions for X: simulation results for lo_hipK after 100 iterations (Figure 13) simulation results for hi_hipK after 600 iterations (Figure 14) For both results, the model was simulated and allowed to run until a stable trajectory was achieved. This stable cycle was used to compare with the original kinematic data. Finally, a disturbance was added to the simulation (results below).
	Overlay Movies
Figure 15 Overlay for lo_hipK (movie) Figure 16 Overlay for hi_hipK (movie)	Figures 15 and 16 show the kinematic data and the simulation results for both optimization solutions. As can be seen, both trajectories are qualitatively similar. The trajectory for hi_hipK appears to be out of phase.
	Trajectory Comparison
Figure 17 X position for simulated trajectory of lo_hipK Figure 18 Y position for simulated trajectory of lo_hipK Figure 19 Orientation for simulated trajectory of lo_hipK Figure 20 X position for simulated trajectory of hi_hipK Figure 21 Y position for simulated trajectory of hi_hipK Figure 22 Orientation for simulated trajectory of hi_hipK	These figures compare the kinematic data with one stride period of the simulations (taken after the simulation has converged to a stable trajectory). As seen in the figures, the results for lo_hipK appear to match the position trajectories slightly better than the results for hi_hipK. In particular, note that the simulation for hi_hipK in the X direction is out of phase with the kinematic data. Note that the kinematic data is not regular (non-cyclic), while the simulation results return to the same state at the end of the stride period.
	Disturbance Rejection Movies
Figure 23 Disturbance rejection for lo_hipK Figure 24 Disturbance rejection for hi_hipK	For both solutions, the model was simulated and a large disturbance was introduced during the simulation. The disturbance consisted of the addition of a velocity in the x direction of 1m/s. Figures 23 and 24 show the resulting motion. As seen from the movies, the solution for hi_hipK is more robust to the disturbance. The solution for lo_hipK takes about three of four stride cycles to converge back to the particular direction.