Graphing Dogleg vs LM with beta distribution error bars

cython/gtsam/tests/test_DogLegOptimizer.py

@@ -9,17 +9,21 @@ DoglegOptimizer unit tests.
 Author: Frank Dellaert
 """
 # pylint: disable=no-member, invalid-name
+
+import math
 import unittest
 
 import gtsam
+import matplotlib.pyplot as plt
 import numpy as np
 from gtsam.utils.test_case import GtsamTestCase
 
 
 class TestDoglegOptimizer(GtsamTestCase):
+    """Test Dogleg vs LM, isnpired by issue #452."""
 
     def test_DoglegOptimizer(self):
-        # Linearization point
+        # Ground truth solution
         T11 = gtsam.Pose2(0, 0, 0)
         T12 = gtsam.Pose2(1, 0, 0)
         T21 = gtsam.Pose2(0, 1, 0)
@@ -34,7 +38,7 @@ class TestDoglegOptimizer(GtsamTestCase):
         graph.add(gtsam.PriorFactorPose2(21, T21, prior))
 
         # Odometry
-        model = gtsam.noiseModel_Diagonal.Sigmas(np.array([0.01, 0.01, 1e6]))
+        model = gtsam.noiseModel_Diagonal.Sigmas(np.array([0.01, 0.01, 0.3]))
         graph.add(gtsam.BetweenFactorPose2(11, 12, T11.between(T12), model))
         graph.add(gtsam.BetweenFactorPose2(21, 22, T21.between(T22), model))
 
@@ -42,36 +46,63 @@ class TestDoglegOptimizer(GtsamTestCase):
         model_rho = gtsam.noiseModel_Isotropic.Sigma(1, 0.01)
         graph.add(gtsam.RangeFactorPose2(12, 22, 1.0, model_rho))
 
-        # Print graph
-        print(graph)
+        num_samples = 1000
+        print("num samples = {}%".format(num_samples))
 
-        sigma = 0.1
-        values = gtsam.Values()
-        values.insert(11, T11.retract(np.random.normal(0, sigma, 3)))
-        values.insert(12, T12)
-        values.insert(21, T21)
-        values.insert(22, T22)
-        linearized = graph.linearize(values)
+        params = gtsam.DoglegParams()
+        params.setDeltaInitial(10)  # default was 1.0
 
-        # Get Jacobian
-        ordering = gtsam.Ordering()
-        ordering.push_back(11)
-        ordering.push_back(21)
-        ordering.push_back(12)
-        ordering.push_back(22)
-        A, b = linearized.jacobian(ordering)
-        Q = np.dot(A.transpose(), A)
-        print(np.linalg.det(Q))
+        # Add progressively more noise to ground truth
+        sigmas = [0.01, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20]
+        n = len(sigmas)
+        p_dl, s_dl, p_lm, s_lm = [0]*n, [0]*n, [0]*n, [0]*n
+        for i, sigma in enumerate(sigmas):
+            dl_fails, lm_fails = 0, 0
+            # Attempt num_samples optimizations for both DL and LM
+            for _attempt in range(num_samples):
+                initial = gtsam.Values()
+                initial.insert(11, T11.retract(np.random.normal(0, sigma, 3)))
+                initial.insert(12, T12.retract(np.random.normal(0, sigma, 3)))
+                initial.insert(21, T21.retract(np.random.normal(0, sigma, 3)))
+                initial.insert(22, T22.retract(np.random.normal(0, sigma, 3)))
 
-        bn = linearized.eliminateSequential(ordering)
+                # Run dogleg optimizer
+                dl = gtsam.DoglegOptimizer(graph, initial, params)
+                result = dl.optimize()
+                dl_fails += graph.error(result) > 1e-9
 
-        # Print gradient
-        linearized.gradientAtZero()
+                # Run
+                lm = gtsam.LevenbergMarquardtOptimizer(graph, initial)
+                result = lm.optimize()
+                lm_fails += graph.error(result) > 1e-9
 
-        # Run dogleg optimizer
-        dl = gtsam.DoglegOptimizer(graph, values)
-        result = dl.optimize()
-        print(graph.error(result))
+            # Calculate Bayes estimate of success probability
+            # using a beta prior of alpha=0.5, beta=0.5
+            alpha, beta = 0.5, 0.5
+            v = num_samples+alpha+beta
+            p_dl[i] = (num_samples-dl_fails+alpha)/v
+            p_lm[i] = (num_samples-lm_fails+alpha)/v
+
+            def stddev(p):
+                """Calculate standard deviation."""
+                return math.sqrt(p*(1-p)/(1+v))
+
+            s_dl[i] = stddev(p_dl[i])
+            s_lm[i] = stddev(p_lm[i])
+
+            fmt = "sigma= {}:\tDL success {:.2f}% +/- {:.2f}%, LM success {:.2f}% +/- {:.2f}%"
+            print(fmt.format(sigma,
+                             100*p_dl[i], 100*s_dl[i],
+                             100*p_lm[i], 100*s_lm[i]))
+
+        fig, ax = plt.subplots()
+        dl_plot = plt.errorbar(sigmas, p_dl, yerr=s_dl, label="Dogleg")
+        lm_plot = plt.errorbar(sigmas, p_lm, yerr=s_lm, label="LM")
+        plt.title("Dogleg emprical success vs. LM")
+        plt.legend(handles=[dl_plot, lm_plot])
+        ax.set_xlim(0, sigmas[-1]+1)
+        ax.set_ylim(0, 1)
+        plt.show()
 
 
 if __name__ == "__main__":