def test_constructor(self):
"""Test that a ``PoseR2`` instance can be created.
"""
r2a = PoseR2([1, 2])
r2b = PoseR2(np.array([3, 4]))
self.assertIsInstance(r2a, PoseR2)
self.assertIsInstance(r2b, PoseR2)
def test_sub(self):
"""Test that the overloaded ``__sub__`` method works as expected.
"""
r2a = PoseR2([1, 2])
r2b = PoseR2([3, 4])
expected = PoseR2([2, 2])
self.assertAlmostEqual(
np.linalg.norm((r2b - r2a).to_array() - expected), 0.)
def test_add(self):
"""Test that the overloaded ``__add__`` method works as expected.
"""
r2a = PoseR2([1, 2])
r2b = PoseR2([3, 4])
expected = PoseR2([4, 6])
self.assertAlmostEqual(
np.linalg.norm((r2a + r2b).to_array() - expected), 0.)
r2a += r2b
self.assertAlmostEqual(np.linalg.norm(r2a.to_array() - expected), 0.)
def test_orientation(self):
"""Test that the ``orientation`` property works as expected.
"""
r2 = PoseR2([1, 2])
self.assertEqual(r2.orientation, 0.)
def test_calc_jacobians(self):
"""Test that the ``calc_jacobians`` method works as expected.
"""
p1 = PoseR2([1, 2])
p2 = PoseR2([3, 4])
estimate = PoseR2([0, 0])
v1 = Vertex(1, p1)
v2 = Vertex(2, p2)
e = EdgeOdometry([1, 2], np.eye(2), estimate, [v1, v2])
jacobians = e.calc_jacobians()
self.assertAlmostEqual(np.linalg.norm(jacobians[0] - np.eye(2)), 0.)
self.assertAlmostEqual(np.linalg.norm(jacobians[1] + np.eye(2)), 0.)
def test_copy(self):
"""Test that the ``copy`` method works as expected.
"""
p1 = PoseR2([1, 2])
p2 = p1.copy()
p2[0] = 0
self.assertEqual(p1[0], 1)
def test_inverse(self):
"""Test that the ``inverse`` property works as expected.
"""
r2 = PoseR2([1, 2])
true_inv = np.array([-1, -2])
self.assertAlmostEqual(
np.linalg.norm(r2.inverse.to_array() - true_inv), 0.)
def test_to_compact(self):
"""Test that the ``to_compact`` method works as expected.
"""
r2 = PoseR2([1, 2])
arr = r2.to_compact()
self.assertIsInstance(arr, np.ndarray)
self.assertNotIsInstance(arr, PoseR2)
self.assertAlmostEqual(np.linalg.norm(arr - np.array([1, 2])), 0.)
def setUp(self):
r"""Setup a simple ``Graph`` in :math:`\mathbb{R}^2`.
"""
np.random.seed(0)
p1 = PoseR2(np.random.random_sample(2))
p2 = PoseR2(np.random.random_sample(2))
p3 = PoseR2(np.random.random_sample(2))
estimate = PoseR2([0, 0])
v1 = Vertex(1, p1)
v2 = Vertex(2, p2)
v3 = Vertex(3, p3)
e1 = EdgeOdometry([1, 2], np.eye(2), estimate, [v1, v2])
e2 = EdgeOdometry([3, 2], 2 * np.eye(2), estimate, [v3, v2])
self.g = Graph([e1, e2], [v1, v2, v3])
def test_position(self):
"""Test that the ``position`` property works as expected.
"""
r2 = PoseR2([1, 2])
pos = r2.position
true_pos = np.array([1, 2])
self.assertIsInstance(pos, np.ndarray)
self.assertNotIsInstance(pos, PoseR2)
self.assertAlmostEqual(np.linalg.norm(true_pos - pos), 0.)
def test_jacobian_self_ominus_other_compact(self):
"""Test that the ``jacobian_self_ominus_other_wrt_self_compact`` and ``jacobian_self_ominus_other_wrt_other_compact`` methods are correctly implemented.
"""
np.random.seed(0)
for _ in range(10):
p1 = PoseR2(np.random.random_sample(2))
p2 = PoseR2(np.random.random_sample(2))
v1 = Vertex(1, p1)
v2 = Vertex(2, p2)
e = EdgeOMinusCompact([1, 2], np.eye(2), np.zeros(2), [v1, v2])
numerical_jacobians = BaseEdge.calc_jacobians(e)
analytical_jacobians = e.calc_jacobians()
self.assertEqual(len(numerical_jacobians),
len(analytical_jacobians))
for n, a in zip(numerical_jacobians, analytical_jacobians):
self.assertAlmostEqual(np.linalg.norm(n - a), 0.)
def test_calc_chi2_gradient_hessian(self):
"""Test that the ``calc_chi2_gradient_hessian`` method works as expected.
"""
p1 = PoseR2([1, 3])
p2 = PoseR2([2, 4])
estimate = PoseR2([0, 0])
v1 = Vertex(0, p1, 0)
v2 = Vertex(1, p2, 1)
e = EdgeOdometry([0, 1], np.eye(2), estimate, [v1, v2])
chi2, gradient, hessian = e.calc_chi2_gradient_hessian()
self.assertEqual(chi2, 2.)
self.assertAlmostEqual(np.linalg.norm(gradient[0] + np.ones(2)), 0.)
self.assertAlmostEqual(np.linalg.norm(gradient[1] - np.ones(2)), 0.)
self.assertAlmostEqual(np.linalg.norm(hessian[(0, 0)] - np.eye(2)), 0.)
self.assertAlmostEqual(np.linalg.norm(hessian[(0, 1)] + np.eye(2)), 0.)
self.assertAlmostEqual(np.linalg.norm(hessian[(1, 1)] - np.eye(2)), 0.)
def test_plot(self):
"""Test that the ``plot`` method is not implemented.
"""
v_none = Vertex(0, None)
v_r2 = Vertex(1, PoseR2([1, 2]))
v_se2 = Vertex(2, PoseSE2([1, 2], 3))
v_r3 = Vertex(3, PoseR3([1, 2, 3]))
v_se3 = Vertex(4, PoseSE3([1, 2, 3], [0.5, 0.5, 0.5, 0.5]))
with self.assertRaises(NotImplementedError):
e = EdgeOdometry(0, 1, 0, [v_none, v_none])
e.plot()
for v in [v_r2, v_se2, v_r3, v_se3]:
e = EdgeOdometry(0, 1, 0, [v, v])
e.plot()
def test_plot(self):
"""Test that a ``Vertex`` can be plotted.
"""
v_none = Vertex(0, None)
v_r2 = Vertex(1, PoseR2([1, 2]))
v_se2 = Vertex(2, PoseSE2([1, 2], 3))
v_r3 = Vertex(3, PoseR3([1, 2, 3]))
v_se3 = Vertex(4, PoseSE3([1, 2, 3], [0.5, 0.5, 0.5, 0.5]))
with self.assertRaises(NotImplementedError):
v_none.plot()
for v in [v_r2, v_se2, v_r3, v_se3]:
fig = plt.figure()
if len(v.pose.position) == 3:
fig.add_subplot(111, projection='3d')
v.plot()