/// <summary>
/// Simple overall prediction
/// </summary>
/// <param name="scene"></param>
/// <param name="u"></param>
/// <param name="delta_t"></param>
public void predict_filter_slow (Scene_Single scene, Vector u, float delta_t)
{
Debug.WriteLine("*** SLOW PREDICTION ***");
// What we need to do for the prediction:
// Calculate f and grad_f_x
// Calculate Q
// Form x(k+1|k) and P(k+1|k)
int size = (int)scene.get_total_state_size();
// First form original total state and covariance
Vector x = new Vector(size);
MatrixFixed P = new MatrixFixed(size, size);
scene.construct_total_state_and_covariance(ref x, ref P);
// Make model calculations: store results in RES matrices
Vector xv = scene.get_xv();
//Vector xv = new Vector(scene.get_xv());
scene.get_motion_model().func_fv_and_dfv_by_dxv(xv, u, delta_t);
scene.get_motion_model().func_Q(scene.get_xv(), u, delta_t);
// Find new state f
Vector f = new Vector(size);
// Feature elements of f are the same as x
f.Update(x);
f.Update(scene.get_motion_model().get_fvRES(), 0);
// Find new P
// Since most elements of df_by_dx are zero...
MatrixFixed df_by_dx = new MatrixFixed(size, size);
df_by_dx.Fill(0.0f);
// Fill the rest of the elements of df_by_dx: 1 on diagonal for features
for (int i = (int)scene.get_motion_model().STATE_SIZE; i < df_by_dx.Rows; i++)
df_by_dx[i,i] = 1.0f;
df_by_dx.Update(scene.get_motion_model().get_dfv_by_dxvRES(), 0, 0);
// Calculate the process noise
MatrixFixed Q = new MatrixFixed(size, size);
Q.Fill(0.0f);
Q.Update(scene.get_motion_model().get_QxRES(), 0, 0);
P.Update(df_by_dx * P * df_by_dx.Transpose());
P += Q;
scene.fill_state_and_covariance(f, P);
}
/// <summary>
/// Constructor for partially-initialised features.
/// </summary>
/// <param name="id">reference to the feature</param>
/// <param name="lab"></param>
/// <param name="list_pos"></param>
/// <param name="scene"></param>
/// <param name="h"></param>
/// <param name="p_i_f_m_m"></param>
public Feature(classimage_mono id, uint lab, uint list_pos,
Scene_Single scene, Vector h,
Partially_Initialised_Feature_Measurement_Model p_i_f_m_m,
Vector feature_colour)
{
feature_measurement_model = p_i_f_m_m;
partially_initialised_feature_measurement_model = p_i_f_m_m;
fully_initialised_feature_measurement_model = null;
// Stuff below substituted from Feature_common
// Feature_common(id, lab, list_pos, scene, h);
feature_constructor_bookeeping();
identifier = id;
label = lab;
position_in_list = list_pos; // Position of new feature in list
position_in_total_state_vector = 0; // This should be set properly
colour = feature_colour;
//when feature is added
// Save the vehicle position where this feature was acquired
scene.get_motion_model().func_xp(scene.get_xv());
//xp_orig = scene.get_motion_model().get_xpRES();
xp_orig = new Vector(scene.get_motion_model().get_xpRES());
// Call model functions to calculate feature state, measurement noise
// and associated Jacobians. Results are stored in RES matrices
// First calculate "position state" and Jacobian
scene.get_motion_model().func_xp(scene.get_xv());
scene.get_motion_model().func_dxp_by_dxv(scene.get_xv());
// Now ask the model to initialise the state vector and calculate Jacobians
// so that I can go and calculate the covariance matrices
partially_initialised_feature_measurement_model.func_ypi_and_dypi_by_dxp_and_dypi_by_dhi_and_Ri(h, scene.get_motion_model().get_xpRES());
// State y
//y = partially_initialised_feature_measurement_model.get_ypiRES();
y = new Vector(partially_initialised_feature_measurement_model.get_ypiRES());
// Temp_FS1 will store dypi_by_dxv
MatrixFixed Temp_FS1 =
partially_initialised_feature_measurement_model.get_dypi_by_dxpRES() *
scene.get_motion_model().get_dxp_by_dxvRES();
// Pxy
Pxy = scene.get_Pxx() * Temp_FS1.Transpose();
// Pyy
Pyy = Temp_FS1 * scene.get_Pxx() * Temp_FS1.Transpose()
+ partially_initialised_feature_measurement_model.get_dypi_by_dhiRES()
* partially_initialised_feature_measurement_model.get_RiRES()
* partially_initialised_feature_measurement_model.get_dypi_by_dhiRES().Transpose();
// Covariances of this feature with others
int j = 0;
foreach (Feature it in scene.get_feature_list_noconst())
{
if (j < position_in_list)
{
// new Pypiyj = dypi_by_dxv . Pxyj
// Size of this is FEATURE_STATE_SIZE(new) by FEATURE_STATE_SIZE(old)
MatrixFixed m = it.get_Pxy();
MatrixFixed newPyjypi_to_store = (Temp_FS1 * m).Transpose();
//add to the list
matrix_block_list.Add(newPyjypi_to_store);
}
j++;
}
known_feature_label = -1;
}
// Function to find a region in an image guided by current motion prediction
// which doesn't overlap existing features
public static bool FindNonOverlappingRegion(Scene_Single scene,
Vector local_u,
float delta_t,
Partially_Initialised_Feature_Measurement_Model default_feature_type_for_initialisation,
uint camera_width,
uint camera_height,
uint BOXSIZE,
ref int init_feature_search_ustart,
ref int init_feature_search_vstart,
ref int init_feature_search_ufinish,
ref int init_feature_search_vfinish,
uint FEATURE_INIT_STEPS_TO_PREDICT,
float FEATURE_INIT_DEPTH_HYPOTHESIS,
Random rnd)
{
ThreeD_Motion_Model threed_motion_model = (ThreeD_Motion_Model)scene.get_motion_model();
Vector local_xv = new Vector(scene.get_xv());
for (uint i = 0; i < FEATURE_INIT_STEPS_TO_PREDICT; i++)
{
scene.get_motion_model().func_fv_and_dfv_by_dxv(local_xv, local_u, delta_t);
local_xv.Update(scene.get_motion_model().get_fvRES());
}
threed_motion_model.func_xp(local_xv);
Vector local_xp = new Vector(threed_motion_model.get_xpRES());
threed_motion_model.func_r(local_xp);
Vector3D rW = threed_motion_model.get_rRES();
threed_motion_model.func_q(local_xp);
Quaternion qWR = threed_motion_model.get_qRES();
// yW = rW + RWR hR
Vector3D hR = new Vector3D(0.0f, 0.0f, FEATURE_INIT_DEPTH_HYPOTHESIS);
// Used Inverse + transpose because this was't compiling the normal way
Vector3D yW = new Vector3D(rW + qWR.RotationMatrix() * hR);
// Then project that point into the current camera position
scene.get_motion_model().func_xp(scene.get_xv());
Fully_Initialised_Feature_Measurement_Model fully_init_fmm =
(Fully_Initialised_Feature_Measurement_Model)(default_feature_type_for_initialisation.more_initialised_feature_measurement_model);
Vector yWVNL = yW.GetVNL3();
fully_init_fmm.func_hi_and_dhi_by_dxp_and_dhi_by_dyi(yWVNL, scene.get_motion_model().get_xpRES());
// Now, this defines roughly how much we expect a feature initialised to move
float suggested_u = fully_init_fmm.get_hiRES()[0];
float suggested_v = fully_init_fmm.get_hiRES()[1];
float predicted_motion_u = camera_width / 2.0f - suggested_u;
float predicted_motion_v = camera_height / 2.0f - suggested_v;
// So, the limits of a "safe" region within which we can initialise
// features so that they end up staying within the screen
// (Making the approximation that the whole screen moves like the
// centre point)
int safe_feature_search_ustart = (int)(-predicted_motion_u);
int safe_feature_search_vstart = (int)(-predicted_motion_v);
int safe_feature_search_ufinish = (int)(camera_width - predicted_motion_u);
int safe_feature_search_vfinish = (int)(camera_height - predicted_motion_v);
if (safe_feature_search_ustart < ((int)((BOXSIZE - 1) / 2) + 1))
safe_feature_search_ustart = (int)((BOXSIZE - 1) / 2 + 1);
if (safe_feature_search_ufinish > (int)camera_width - ((int)((BOXSIZE - 1) / 2) + 1))
safe_feature_search_ufinish = (int)(camera_width - (BOXSIZE - 1) / 2 - 1);
if (safe_feature_search_vstart < ((int)((BOXSIZE - 1) / 2) + 1))
safe_feature_search_vstart = (int)((BOXSIZE - 1) / 2 + 1);
if (safe_feature_search_vfinish > (int)camera_height - ((int)((BOXSIZE - 1) / 2) + 1))
safe_feature_search_vfinish = (int)(camera_height - (BOXSIZE - 1) / 2 - 1);
return FindNonOverlappingRegionNoPredict(safe_feature_search_ustart,
safe_feature_search_vstart,
safe_feature_search_ufinish,
safe_feature_search_vfinish,
scene,
ref init_feature_search_ustart,
ref init_feature_search_vstart,
ref init_feature_search_ufinish,
ref init_feature_search_vfinish, rnd);
}
/// <summary>
/// Function to find non-overlapping region over without prediction
/// this is really the service function called by both the above
/// </summary>
/// <param name="safe_feature_search_ustart"></param>
/// <param name="safe_feature_search_vstart"></param>
/// <param name="safe_feature_search_ufinish"></param>
/// <param name="safe_feature_search_vfinish"></param>
/// <param name="scene"></param>
/// <param name="init_feature_search_ustart"></param>
/// <param name="init_feature_search_vstart"></param>
/// <param name="init_feature_search_ufinish"></param>
/// <param name="init_feature_search_vfinish"></param>
/// <param name="rnd"></param>
/// <returns></returns>
public static bool FindNonOverlappingRegionNoPredict(
int safe_feature_search_ustart,
int safe_feature_search_vstart,
int safe_feature_search_ufinish,
int safe_feature_search_vfinish,
Scene_Single scene,
ref int init_feature_search_ustart,
ref int init_feature_search_vstart,
ref int init_feature_search_ufinish,
ref int init_feature_search_vfinish,
Random rnd)
{
int i, j;
//if (Camera_Constants.DEBUGDUMP) cout << "FNOLRNP timer start: " << timerlocal << endl;
int INIT_FEATURE_SEARCH_WIDTH = 80;
int INIT_FEATURE_SEARCH_HEIGHT = 60;
// Within this, choose a random region
// Check that we've got some room for manouevre
if ((safe_feature_search_ufinish - safe_feature_search_ustart > INIT_FEATURE_SEARCH_WIDTH) &&
(safe_feature_search_vfinish - safe_feature_search_vstart > INIT_FEATURE_SEARCH_HEIGHT))
{
// Try a few times to get one that's not overlapping with any features
// we know about
int NUMBER_OF_RANDOM_INIT_FEATURE_SEARCH_REGION_TRIES = 5;
int FEATURE_SEPARATION_MINIMUM = (int)Camera_Constants.BOXSIZE*3;
// Build vectors of feature positions so we only have to work them out once
ArrayList u_array = new ArrayList();
ArrayList v_array = new ArrayList();
Feature it;
for (j = 0; j < scene.get_feature_list().Count; j++)
{
it = (Feature)(scene.get_feature_list())[j];
//Vector z = it.get_z();
//u_array.Add(z[0]);
//v_array.Add(z[1]);
if (it.get_feature_measurement_model().fully_initialised_flag)
{
//Vector z = it.get_z();
//u_array.Add(z[0]);
//v_array.Add(z[1]);
Fully_Initialised_Feature_Measurement_Model fifmm =
(Fully_Initialised_Feature_Measurement_Model)(it.get_feature_measurement_model());
fifmm.func_hi_and_dhi_by_dxp_and_dhi_by_dyi(it.get_y(), scene.get_motion_model().get_xpRES());
// Check that this is not a feature behind the camera
if (it.get_feature_measurement_model().feature_graphics_type == "THREED_POINT")
{
fifmm.func_zeroedyigraphics_and_Pzeroedyigraphics(it.get_y(),
scene.get_xv(), scene.get_Pxx(), it.get_Pxy(), it.get_Pyy());
if (fifmm.get_zeroedyigraphicsRES()[2] > 0)
{
u_array.Add(fifmm.get_hiRES()[0]);
v_array.Add(fifmm.get_hiRES()[1]);
}
}
}
}
//if (Camera_Constants.DEBUGDUMP) cout << "FNOLRNP timer after functions: " << timerlocal << endl;
bool feature_found = false;
i = 0;
while (i < NUMBER_OF_RANDOM_INIT_FEATURE_SEARCH_REGION_TRIES)
{
int u_offset = (int)((safe_feature_search_ufinish - safe_feature_search_ustart
- INIT_FEATURE_SEARCH_WIDTH) * (rnd.Next(10000) / 10000.0f));
int v_offset = (int)((safe_feature_search_vfinish - safe_feature_search_vstart
- INIT_FEATURE_SEARCH_HEIGHT) * (rnd.Next(10000) / 10000.0f));
init_feature_search_ustart = safe_feature_search_ustart + u_offset;
init_feature_search_ufinish = init_feature_search_ustart + INIT_FEATURE_SEARCH_WIDTH;
init_feature_search_vstart = safe_feature_search_vstart + v_offset;
init_feature_search_vfinish = init_feature_search_vstart + INIT_FEATURE_SEARCH_HEIGHT;
bool found_a_feature_in_region_flag = false;
// These arrays will be the same size
float uit, vit;
for (j = 0; j < u_array.Count; j++)
{
uit = (float)u_array[j];
vit = (float)v_array[j];
if ((uit >= init_feature_search_ustart - FEATURE_SEPARATION_MINIMUM) &&
(uit < init_feature_search_ufinish + FEATURE_SEPARATION_MINIMUM) &&
(vit >= init_feature_search_vstart - FEATURE_SEPARATION_MINIMUM) &&
(vit < init_feature_search_vfinish + FEATURE_SEPARATION_MINIMUM))
{
found_a_feature_in_region_flag = true;
feature_found = true;
break;
}
}
if (!found_a_feature_in_region_flag) break;
i++;
}
if (!feature_found)
{
return false;
}
}
else
{
return false;
}
return true;
}
/// <summary>
/// Fast version of predict filter
/// </summary>
/// <param name="scene"></param>
/// <param name="u">Control vector of accelerations</param>
/// <param name="delta_t">time step in seconds</param>
public void predict_filter_fast(Scene_Single scene, Vector u, float delta_t)
{
Vector xv = scene.get_xv();
// Make model calculations: results are stored in RES matrices
scene.get_motion_model().func_fv_and_dfv_by_dxv(xv, u, delta_t);
scene.get_motion_model().func_Q(xv, u, delta_t);
Vector fvRES = scene.get_motion_model().get_fvRES();
xv.Update(fvRES);
Motion_Model mm = scene.get_motion_model();
MatrixFixed Pxx = scene.get_Pxx();
MatrixFixed dfv_by_dxvRES = mm.get_dfv_by_dxvRES();
MatrixFixed dfv_by_dxvRES_transposed = dfv_by_dxvRES.Transpose();
MatrixFixed QxRES = mm.get_QxRES();
Pxx.Update((dfv_by_dxvRES * Pxx * dfv_by_dxvRES_transposed) + QxRES);
// Change the covariances between vehicle state and feature states
foreach (Feature it in scene.feature_list)
{
MatrixFixed Pxy = it.get_Pxy();
Pxy.Update(dfv_by_dxvRES * Pxy);
}
}