mirror of
https://github.com/WallabyLester/RBF-aPID-Controller.git
synced 2026-05-11 05:39:16 +08:00
Andru Liu
37 lines
1.3 KiB
Python
37 lines
1.3 KiB
Python
import unittest
|
|
import numpy as np
|
|
|
|
from RBF_numpy import RBFNetwork
|
|
|
|
class TestRBFNetwork(unittest.TestCase):
|
|
def setUp(self):
|
|
"""Set up an RBFNetwork instance for testing."""
|
|
self.input_dim = 3
|
|
self.n_centers = 5
|
|
self.x = np.array([6.0, 0.5, 0.2])
|
|
self.rbf_network = RBFNetwork(self.input_dim, self.n_centers)
|
|
|
|
def test_gaussian(self):
|
|
"""Test the Gaussian function."""
|
|
center = np.random.rand(1, self.input_dim)
|
|
expected_output = np.exp(-np.linalg.norm(self.x - center) ** 2 / (2 * self.rbf_network.sigma ** 2))
|
|
output = self.rbf_network.gaussian(self.x, center)
|
|
self.assertAlmostEqual(output, expected_output, places=5)
|
|
|
|
def test_predict(self):
|
|
"""Test the predict function."""
|
|
output_before = self.rbf_network.predict(self.x)
|
|
self.assertIsInstance(output_before, float)
|
|
[self.assertNotAlmostEqual(self.rbf_network.weights[i], self.rbf_network.weights[i+1])
|
|
for i in range(len(self.rbf_network.weights)-1)]
|
|
|
|
target = 1.0
|
|
self.rbf_network.train(self.x, target)
|
|
output_after = self.rbf_network.predict(self.x)
|
|
self.assertIsInstance(output_after, float)
|
|
self.assertNotEqual(output_before, output_after)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|