From 79f7861f0c14a0c3a502be0ff3447a1dc737ea28 Mon Sep 17 00:00:00 2001 From: jerredchen Date: Sun, 24 Oct 2021 15:34:49 -0400 Subject: [PATCH] made changes according to Frank --- python/gtsam/examples/Pose2ISAM2Example.py | 158 ++++++++++-------- python/gtsam/examples/Pose3ISAM2Example.py | 182 ++++++++++++--------- 2 files changed, 189 insertions(+), 151 deletions(-) diff --git a/python/gtsam/examples/Pose2ISAM2Example.py b/python/gtsam/examples/Pose2ISAM2Example.py index 5336bc34e..fac7ecc1a 100644 --- a/python/gtsam/examples/Pose2ISAM2Example.py +++ b/python/gtsam/examples/Pose2ISAM2Example.py @@ -13,23 +13,28 @@ Modelled after: - Pose2SLAMExample by: Alex Cunningham (C++), Kevin Deng & Frank Dellaert (Python) """ -from __future__ import print_function import math import numpy as np -from numpy.random import multivariate_normal as mult_gauss import gtsam import matplotlib.pyplot as plt import gtsam.utils.plot as gtsam_plot -def Pose2SLAM_ISAM2_plot(graph, current_estimate): - """Plots incremental state of the robot for 2D Pose SLAM using iSAM2 +def report_on_progress(graph: gtsam.NonlinearFactorGraph, current_estimate: gtsam.Values, + key: int): + """Print and plot incremental progress of the robot for 2D Pose SLAM using iSAM2. Based on version by: - Ellon Paiva (Python), - Duy Nguyen Ta and Frank Dellaert (MATLAB) """ + # Print the current estimates computed using iSAM2. + print("*"*50 + f"\nInference after State {key+1}:\n") + print(current_estimate) + + # Compute the marginals for all states in the graph. marginals = gtsam.Marginals(graph, current_estimate) + # Plot the newly updated iSAM2 inference. fig = plt.figure(0) axes = fig.gca() plt.cla() @@ -43,19 +48,47 @@ def Pose2SLAM_ISAM2_plot(graph, current_estimate): axes.set_xlim(-1, 5) axes.set_ylim(-1, 3) plt.pause(1) + return marginals +def vector3(x, y, z): + """Create a 3D double numpy array.""" + return np.array([x, y, z], dtype=float) + +def determine_loop_closure(odom: np.ndarray, current_estimate: gtsam.Values, + key: int, xy_tol=0.6, theta_tol=0.3) -> int: + """Simple brute force approach which iterates through previous states + and checks for loop closure. + + Args: + odom: Vector representing noisy odometry (x, y, theta) measurement in the body frame. + current_estimate: The current estimates computed by iSAM2. + key: Key corresponding to the current state estimate of the robot. + xy_tol: Optional argument for the x-y measurement tolerance. + theta_tol: Optional argument for the theta measurement tolerance. + Returns: + k: The key of the state which is helping add the loop closure constraint. + If loop closure is not found, then None is returned. + """ + if current_estimate: + prev_est = current_estimate.atPose2(key+1) + rotated_odom = prev_est.rotation().matrix() @ odom[:2] + curr_xy = np.array([prev_est.x() + rotated_odom[0], + prev_est.y() + rotated_odom[1]]) + curr_theta = prev_est.theta() + odom[2] + for k in range(1, key+1): + pose_xy = np.array([current_estimate.atPose2(k).x(), + current_estimate.atPose2(k).y()]) + pose_theta = current_estimate.atPose2(k).theta() + if (abs(pose_xy - curr_xy) <= xy_tol).all() and \ + (abs(pose_theta - curr_theta) <= theta_tol): + return k def Pose2SLAM_ISAM2_example(): """Perform 2D SLAM given the ground truth changes in pose as well as - simple loop closure detection. - """ + simple loop closure detection.""" plt.ion() - def vector3(x, y, z): - """Create 3d double numpy array.""" - return np.array([x, y, z], dtype=float) - # Although this example only uses linear measurements and Gaussian noise models, it is important # to note that iSAM2 can be utilized to its full potential during nonlinear optimization. This example # simply showcases how iSAM2 may be applied to a Pose2 SLAM problem. @@ -66,85 +99,68 @@ def Pose2SLAM_ISAM2_example(): graph = gtsam.NonlinearFactorGraph() initial_estimate = gtsam.Values() - # iSAM2 parameters which can adjust the threshold necessary to force relinearization and how many + # Create iSAM2 parameters which can adjust the threshold necessary to force relinearization and how many # update calls are required to perform the relinearization. parameters = gtsam.ISAM2Params() parameters.setRelinearizeThreshold(0.1) parameters.setRelinearizeSkip(1) isam = gtsam.ISAM2(parameters) - # The ground truth odometry measurements (without noise) of the robot during the trajectory. - odom_arr = [(2, 0, 0), - (2, 0, math.pi/2), - (2, 0, math.pi/2), - (2, 0, math.pi/2), - (2, 0, math.pi/2)] + # Create the ground truth odometry measurements of the robot during the trajectory. + true_odometry = [(2, 0, 0), + (2, 0, math.pi/2), + (2, 0, math.pi/2), + (2, 0, math.pi/2), + (2, 0, math.pi/2)] - # The initial estimates for robot poses 2-5. Pose 1 is initialized by the prior. - # To demonstrate iSAM2 incremental optimization, poor initializations were intentionally inserted. - pose_est = [gtsam.Pose2(2.3, 0.1, -0.2), - gtsam.Pose2(4.1, 0.1, math.pi/2), - gtsam.Pose2(4.0, 2.0, math.pi), - gtsam.Pose2(2.1, 2.1, -math.pi/2), - gtsam.Pose2(1.9, -0.2, 0.2)] + # Corrupt the odometry measurements with gaussian noise to create noisy odometry measurements. + odometry_measurements = [np.random.multivariate_normal(true_odom, ODOMETRY_NOISE.covariance()) + for true_odom in true_odometry] + # Add the prior factor to the factor graph, and poorly initialize the prior pose to demonstrate + # iSAM2 incremental optimization. graph.push_back(gtsam.PriorFactorPose2(1, gtsam.Pose2(0, 0, 0), PRIOR_NOISE)) initial_estimate.insert(1, gtsam.Pose2(0.5, 0.0, 0.2)) - def determine_loop_closure(odom: np.ndarray, current_estimate: gtsam.Values, - xy_tol=0.6, theta_tol=0.3) -> int: - """Simple brute force approach which iterates through previous states - and checks for loop closure. + # Initialize the current estimate which is used during the incremental inference loop. + current_estimate = initial_estimate - Args: - odom: Vector representing noisy odometry (x, y, theta) measurement in the body frame. - current_estimate: The current estimates computed by iSAM2. - xy_tol: Optional argument for the x-y measurement tolerance. - theta_tol: Optional argument for the theta measurement tolerance. - Returns: - k: The key of the state which is helping add the loop closure constraint. - If loop closure is not found, then None is returned. - """ - if current_estimate: - prev_est = current_estimate.atPose2(i+1) - rotated_odom = prev_est.rotation().matrix() @ odom[:2] - curr_xy = np.array([prev_est.x() + rotated_odom[0], - prev_est.y() + rotated_odom[1]]) - curr_theta = prev_est.theta() + odom[2] - for k in range(1, i+1): - pose_xy = np.array([current_estimate.atPose2(k).x(), - current_estimate.atPose2(k).y()]) - pose_theta = current_estimate.atPose2(k).theta() - if (abs(pose_xy - curr_xy) <= xy_tol).all() and \ - (abs(pose_theta - curr_theta) <= theta_tol): - return k + for i in range(len(true_odometry)): - current_estimate = None - for i in range(len(odom_arr)): - # The "ground truth" change between poses - odom_x, odom_y, odom_theta = odom_arr[i] - # Odometry measurement that is received by the robot and corrupted by gaussian noise - odom = mult_gauss(odom_arr[i], ODOMETRY_NOISE.covariance()) - # Determine if there is loop closure based on the odometry measurement and the previous estimate of the state - loop = determine_loop_closure(odom, current_estimate) + # Obtain "ground truth" change between the current pose and the previous pose. + true_odom_x, true_odom_y, true_odom_theta = true_odometry[i] + + # Obtain the noisy odometry that is received by the robot and corrupted by gaussian noise. + noisy_odom_x, noisy_odom_y, noisy_odom_theta = odometry_measurements[i] + + # Determine if there is loop closure based on the odometry measurement and the previous estimate of the state. + loop = determine_loop_closure(odometry_measurements[i], current_estimate, i) + + # Add a binary factor in between two existing states if loop closure is detected. + # Otherwise, add a binary factor between a newly observed state and the previous state. + # Note that the true odometry measurement is used in the factor instead of the noisy odometry measurement. + # This is only to maintain the example consistent for each run. In practice, the noisy odometry measurement is used. if loop: - graph.push_back(gtsam.BetweenFactorPose2(i + 1, loop, gtsam.Pose2(odom_x, odom_y, odom_theta), ODOMETRY_NOISE)) + graph.push_back(gtsam.BetweenFactorPose2(i + 1, loop, + gtsam.Pose2(true_odom_x, true_odom_y, true_odom_theta), ODOMETRY_NOISE)) else: - graph.push_back(gtsam.BetweenFactorPose2(i + 1, i + 2, gtsam.Pose2(odom_x, odom_y, odom_theta), ODOMETRY_NOISE)) - initial_estimate.insert(i + 2, pose_est[i]) - # Incremental update to iSAM2's internal Baye's tree, which only optimizes upon the added variables - # as well as any affected variables + graph.push_back(gtsam.BetweenFactorPose2(i + 1, i + 2, + gtsam.Pose2(true_odom_x, true_odom_y, true_odom_theta), ODOMETRY_NOISE)) + + # Compute and insert the initialization estimate for the current pose using the noisy odometry measurement. + computed_estimate = current_estimate.atPose2(i + 1).compose(gtsam.Pose2(noisy_odom_x, noisy_odom_y, noisy_odom_theta)) + initial_estimate.insert(i + 2, computed_estimate) + + # Perform incremental update to iSAM2's internal Bayes tree, optimizing only the affected variables. isam.update(graph, initial_estimate) - # Another iSAM2 update can be performed for additional optimization - isam.update() current_estimate = isam.calculateEstimate() - print("*"*50) - print(f"Inference after State {i+1}:") - print(current_estimate) - marginals = Pose2SLAM_ISAM2_plot(graph, current_estimate) + + # Report all current state estimates from the iSAM2 optimzation. + marginals = report_on_progress(graph, current_estimate, i) initial_estimate.clear() - # Print the final covariance matrix for each pose after completing inference - for i in range(1, len(odom_arr)+1): + + # Print the final covariance matrix for each pose after completing inference on the trajectory. + for i in range(1, len(true_odometry)+1): print(f"X{i} covariance:\n{marginals.marginalCovariance(i)}\n") plt.ioff() diff --git a/python/gtsam/examples/Pose3ISAM2Example.py b/python/gtsam/examples/Pose3ISAM2Example.py index 3fcdcd7ec..3e78339bd 100644 --- a/python/gtsam/examples/Pose3ISAM2Example.py +++ b/python/gtsam/examples/Pose3ISAM2Example.py @@ -12,34 +12,36 @@ Modelled after version by: - VisualISAM2Example by: Duy-Nguyen Ta (C++), Frank Dellaert (Python) - Pose2SLAMExample by: Alex Cunningham (C++), Kevin Deng & Frank Dellaert (Python) """ -from __future__ import print_function + import gtsam -from gtsam import Pose3, Rot3 -from gtsam.symbol_shorthand import X import gtsam.utils.plot as gtsam_plot import numpy as np -from numpy import cos, sin, pi -from numpy.random import multivariate_normal as mult_gauss import matplotlib.pyplot as plt - -def Pose3SLAM_ISAM2_plot(graph, current_estimate): - """Plots incremental state of the robot for 3D Pose SLAM using iSAM2 +def report_on_progress(graph: gtsam.NonlinearFactorGraph, current_estimate: gtsam.Values, + key: int): + """Print and plot incremental progress of the robot for 2D Pose SLAM using iSAM2. Based on version by: - Ellon Paiva (Python), - Duy Nguyen Ta and Frank Dellaert (MATLAB) """ + # Print the current estimates computed using iSAM2. + print("*"*50 + f"\nInference after State {key+1}:\n") + print(current_estimate) + + # Compute the marginals for all states in the graph. marginals = gtsam.Marginals(graph, current_estimate) + # Plot the newly updated iSAM2 inference. fig = plt.figure(0) axes = fig.gca(projection='3d') plt.cla() - i = 0 - while current_estimate.exists(X(i)): - gtsam_plot.plot_pose3(0, current_estimate.atPose3(X(i)), 10, - marginals.marginalCovariance(X(i))) + i = 1 + while current_estimate.exists(i): + gtsam_plot.plot_pose3(0, current_estimate.atPose3(i), 10, + marginals.marginalCovariance(i)) i += 1 axes.set_xlim3d(-30, 45) @@ -49,7 +51,6 @@ def Pose3SLAM_ISAM2_plot(graph, current_estimate): return marginals - def createPoses(): """Creates ground truth poses of the robot.""" P0 = np.array([[1, 0, 0, 0], @@ -60,9 +61,9 @@ def createPoses(): [1, 0, 0, 15], [0, 0, 1, 20], [0, 0, 0, 1]]) - P2 = np.array([[cos(pi/4), 0, sin(pi/4), 30], + P2 = np.array([[np.cos(np.pi/4), 0, np.sin(np.pi/4), 30], [0, 1, 0, 30], - [-sin(pi/4), 0, cos(pi/4), 30], + [-np.sin(np.pi/4), 0, np.cos(np.pi/4), 30], [0, 0, 0, 1]]) P3 = np.array([[0, 1, 0, 30], [0, 0, -1, 0], @@ -74,99 +75,120 @@ def createPoses(): [0, 0, 0, 1]]) P5 = P0[:] - return [Pose3(P0), Pose3(P1), Pose3(P2), Pose3(P3), Pose3(P4), Pose3(P5)] + return [gtsam.Pose3(P0), gtsam.Pose3(P1), gtsam.Pose3(P2), + gtsam.Pose3(P3), gtsam.Pose3(P4), gtsam.Pose3(P5)] + +def vector6(x, y, z, a, b, c): + """Create a 6D double numpy array.""" + return np.array([x, y, z, a, b, c], dtype=float) + +def determine_loop_closure(odom: np.ndarray, current_estimate: gtsam.Values, + key: int, xyz_tol=0.6, rot_tol=0.3) -> int: + """Simple brute force approach which iterates through previous states + and checks for loop closure. + + Args: + odom: Vector representing noisy odometry transformation measurement in the body frame, + where the vector represents [roll, pitch, yaw, x, y, z]. + current_estimate: The current estimates computed by iSAM2. + key: Key corresponding to the current state estimate of the robot. + xyz_tol: Optional argument for the translational tolerance. + rot_tol: Optional argument for the rotational tolerance. + Returns: + k: The key of the state which is helping add the loop closure constraint. + If loop closure is not found, then None is returned. + """ + if current_estimate: + rot = gtsam.Rot3.RzRyRx(odom[:3]) + odom_tf = gtsam.Pose3(rot, odom[3:6].reshape(-1,1)) + prev_est = current_estimate.atPose3(key+1) + curr_est = prev_est.transformPoseFrom(odom_tf) + for k in range(1, key+1): + pose = current_estimate.atPose3(k) + if (abs(pose.matrix()[:3,:3] - curr_est.matrix()[:3,:3]) <= rot_tol).all() and \ + (abs(pose.matrix()[:3,3] - curr_est.matrix()[:3,3]) <= xyz_tol).all(): + return k def Pose3_ISAM2_example(): """Perform 3D SLAM given ground truth poses as well as simple loop closure detection.""" plt.ion() - def vector6(x, y, z, a, b, c): - """Create 6d double numpy array.""" - return np.array([x, y, z, a, b, c], dtype=float) - # Although this example only uses linear measurements and Gaussian noise models, it is important # to note that iSAM2 can be utilized to its full potential during nonlinear optimization. This example # simply showcases how iSAM2 may be applied to a Pose2 SLAM problem. PRIOR_NOISE = gtsam.noiseModel.Diagonal.Sigmas(vector6(0.1, 0.1, 0.1, 0.3, 0.3, 0.3)) ODOMETRY_NOISE = gtsam.noiseModel.Diagonal.Sigmas(vector6(0.1, 0.1, 0.1, 0.2, 0.2, 0.2)) - # Create the ground truth poses of the robot trajectory. - poses = createPoses() + # Create a Nonlinear factor graph as well as the data structure to hold state estimates. + graph = gtsam.NonlinearFactorGraph() + initial_estimate = gtsam.Values() - # iSAM2 parameters which can adjust the threshold necessary to force relinearization and how many + # Create iSAM2 parameters which can adjust the threshold necessary to force relinearization and how many # update calls are required to perform the relinearization. parameters = gtsam.ISAM2Params() parameters.setRelinearizeThreshold(0.1) parameters.setRelinearizeSkip(1) isam = gtsam.ISAM2(parameters) - # Create a Nonlinear factor graph as well as the data structure to hold state estimates. - graph = gtsam.NonlinearFactorGraph() - initial_estimate = gtsam.Values() + # Create the ground truth poses of the robot trajectory. + true_poses = createPoses() - # Add prior factor to the first pose with intentionally poor initial estimate - graph.push_back(gtsam.PriorFactorPose3(X(0), poses[0], PRIOR_NOISE)) - initial_estimate.insert(X(0), poses[0].compose(gtsam.Pose3( + # Create the ground truth odometry transformations, xyz translations, and roll-pitch-yaw rotations + # between each robot pose in the trajectory. + odometry_tf = [true_poses[i-1].transformPoseTo(true_poses[i]) for i in range(1, len(true_poses))] + odometry_xyz = [(odometry_tf[i].x(), odometry_tf[i].y(), odometry_tf[i].z()) for i in range(len(odometry_tf))] + odometry_rpy = [odometry_tf[i].rotation().rpy() for i in range(len(odometry_tf))] + + # Corrupt the xyz translations and roll-pitch-yaw rotations with gaussian noise to create noisy odometry measurements. + noisy_measurements = [np.random.multivariate_normal(np.hstack((odometry_rpy[i],odometry_xyz[i])), ODOMETRY_NOISE.covariance()) + for i in range(len(odometry_tf))] + + # Add the prior factor to the factor graph, and poorly initialize the prior pose to demonstrate + # iSAM2 incremental optimization. + graph.push_back(gtsam.PriorFactorPose3(1, true_poses[0], PRIOR_NOISE)) + initial_estimate.insert(1, true_poses[0].compose(gtsam.Pose3( gtsam.Rot3.Rodrigues(-0.1, 0.2, 0.25), gtsam.Point3(0.05, -0.10, 0.20)))) - def determine_loop_closure(odom: np.ndarray, current_estimate: gtsam.Values, - xyz_tol=0.6, rot_tol=0.3) -> int: - """Simple brute force approach which iterates through previous states - and checks for loop closure. + # Initialize the current estimate which is used during the incremental inference loop. + current_estimate = initial_estimate + for i in range(len(odometry_tf)): - Args: - odom: Vector representing noisy odometry transformation measurement in the body frame, - where the vector represents [roll, pitch, yaw, x, y, z]. - current_estimate: The current estimates computed by iSAM2. - xyz_tol: Optional argument for the translational tolerance. - rot_tol: Optional argument for the rotational tolerance. - Returns: - k: The key of the state which is helping add the loop closure constraint. - If loop closure is not found, then None is returned. - """ - if current_estimate: - rot = Rot3.RzRyRx(odom[:3]) - odom_tf = Pose3(rot, odom[3:6].reshape(-1,1)) - prev_est = current_estimate.atPose3(X(i-1)) - curr_est = prev_est.transformPoseFrom(odom_tf) - for k in range(i): - pose = current_estimate.atPose3(X(k)) - if (abs(pose.matrix()[:3,:3] - curr_est.matrix()[:3,:3]) <= rot_tol).all() and \ - (abs(pose.matrix()[:3,3] - curr_est.matrix()[:3,3]) <= xyz_tol).all(): - return k + # Obtain "ground truth" transformation between the current pose and the previous pose. + true_odometry = odometry_tf[i] - current_estimate = None - for i in range(1, len(poses)): - # The "ground truth" change between poses - odom_tf = poses[i-1].transformPoseTo(poses[i]) - odom_xyz = odom_tf.x(), odom_tf.y(), odom_tf.z() - odom_rpy = odom_tf.rotation().rpy() - # Odometry measurement that is received by the robot and corrupted by gaussian noise - measurement = mult_gauss(np.hstack((odom_rpy,odom_xyz)), ODOMETRY_NOISE.covariance()) - loop = determine_loop_closure(measurement, current_estimate) - # If loop closure is detected, then adds a constraint between existing states in the factor graph - if loop is not None: - graph.push_back(gtsam.BetweenFactorPose3(X(i-1), X(loop), gtsam.Pose3(odom_tf), ODOMETRY_NOISE)) + # Obtain the noisy translation and rotation that is received by the robot and corrupted by gaussian noise. + noisy_odometry = noisy_measurements[i] + + # Determine if there is loop closure based on the odometry measurement and the previous estimate of the state. + loop = determine_loop_closure(noisy_odometry, current_estimate, i) + + # Add a binary factor in between two existing states if loop closure is detected. + # Otherwise, add a binary factor between a newly observed state and the previous state. + # Note that the true odometry measurement is used in the factor instead of the noisy odometry measurement. + # This is only to maintain the example consistent for each run. In practice, the noisy odometry measurement is used. + if loop: + graph.push_back(gtsam.BetweenFactorPose3(i + 1, loop, true_odometry, ODOMETRY_NOISE)) else: - graph.push_back(gtsam.BetweenFactorPose3(X(i-1), X(i), gtsam.Pose3(odom_tf), ODOMETRY_NOISE)) - # Intentionally insert poor initializations - initial_estimate.insert(X(i), poses[i].compose(gtsam.Pose3( - gtsam.Rot3.Rodrigues(-0.1, 0.2, 0.25), gtsam.Point3(0.05, -0.10, 0.20)))) - # Performs iSAM2 incremental update based on newly added factors + graph.push_back(gtsam.BetweenFactorPose3(i + 1, i + 2, true_odometry, ODOMETRY_NOISE)) + + # Compute and insert the initialization estimate for the current pose using the noisy odometry measurement. + noisy_tf = gtsam.Pose3(gtsam.Rot3.RzRyRx(noisy_odometry[:3]), noisy_odometry[3:6].reshape(-1,1)) + computed_estimate = current_estimate.atPose3(i + 1).compose(noisy_tf) + initial_estimate.insert(i + 2, computed_estimate) + + # Perform incremental update to iSAM2's internal Bayes tree, optimizing only the affected variables. isam.update(graph, initial_estimate) - # Additional iSAM2 optimization - isam.update() current_estimate = isam.calculateEstimate() - print("*"*50) - print(f"Inference after State {i}:\n") - print(current_estimate) - marginals = Pose3SLAM_ISAM2_plot(graph, current_estimate) + + # Report all current state estimates from the iSAM2 optimization. + marginals = report_on_progress(graph, current_estimate, i) initial_estimate.clear() - # Print the final covariance matrix for each pose after completing inference - i = 0 - while current_estimate.exists(X(i)): - print(f"X{i} covariance:\n{marginals.marginalCovariance(X(i))}\n") + + # Print the final covariance matrix for each pose after completing inference. + i = 1 + while current_estimate.exists(i): + print(f"X{i} covariance:\n{marginals.marginalCovariance(i)}\n") i += 1 plt.ioff()