Alex Liu - Autonomy Bootcamp 2023 #105

bantuerfei0 · 2024-09-24T17:50:39Z

Some tasks might've been misunderstood

Possible missing error handling and redundancy

BalajiLeninrajan

This is a good start! I left some comments on your code, feel free to ask question on you thread if you have any.

BalajiLeninrajan · 2024-09-24T21:38:11Z

modules/bootcamp/detect_landing_pad.py


        # Loop over the boxes list and create a list of bounding boxes
-        bounding_boxes = []
+        # BC NOTE: Need for error checking from returned tuple from BoundingBox.create()
+        bounding_boxes = [bounding_box.BoundingBox.create(rectangle)[1] for rectangle in boxes_cpu]


When the creation of a BoundingBox fails, you would want to return an empty list. Doing that with list comprehension will like kinda clunky so I recommend you refactor this into a for loop.

You should be able to find the method definition of BoundingBox.create, it returns a tuple of a bool and a BoundingBox. If the returned boolean is false that means the creation of the boxes failed. You don't have to check for the conditions for failure manually that's done for in the BoundingBox.create method, you will have to implement returning an empty list of any of the calls to the create method fail.

Haha I was wondering what that creation flag was used for

BalajiLeninrajan · 2024-09-24T21:41:50Z

modules/bootcamp/decision_simple_waypoint.py

@@ -70,6 +74,12 @@ def run(

        # Do something based on the report and the state of this class...

+        if report.status == drone_status.DroneStatus.HALTED:
+            if not self.should_land:


This function runs in an infinite loop, the drone will land at the second iteration of the loop. What you want to do is to navigate to the waypoint until you are within the acceptance radius, then create the land command. (Returning the land command will exit out of the loop)

BalajiLeninrajan · 2024-09-24T22:11:39Z

modules/bootcamp/decision_waypoint_landing_pads.py

+        if not landing_pad_locations:
+            return commands.Command.create_null_command()
+        sorted_locations = landing_pad_locations.copy()
+        sorted_locations.sort(


This is actually a decent solution but it would be faster to use min instead (min in python also has a key attr). This will bring this part down from O(nlogn) to O(n).

BalajiLeninrajan · 2024-09-24T22:21:01Z

modules/bootcamp/decision_waypoint_landing_pads.py

+            )
+        )
+        closest_pad: location.Location = sorted_locations[0]
+        rel_x = closest_pad.location_x - reference_location.location_x


try to avoid abbreviating variable names

BalajiLeninrajan · 2024-09-24T22:22:17Z

modules/bootcamp/decision_waypoint_landing_pads.py

+                command = commands.Command.create_land_command()
+            elif self.finished_goal:
+                command = self.create_close_pad_goal(report.position, landing_pad_locations)
+                self.should_land = True  # should land on the next instruction


Set this once you reach the acceptance radius

BalajiLeninrajan · 2024-09-24T22:22:24Z

modules/bootcamp/decision_waypoint_landing_pads.py

+                self.should_land = True  # should land on the next instruction
+            else:
+                command = self.goal
+                self.finished_goal = True


Set this once you reach the acceptance radius

Improved Task 1, 3, and 4 based on feedback

BalajiLeninrajan

I was leaving comments but I think I confused you, mb 😅. You only want to send command to the drone when it is halted, when it is halted, you check if it is within the acceptance radius, if yes then land if no then set on course to the waypoint again.

BalajiLeninrajan · 2024-09-25T21:12:20Z

modules/bootcamp/decision_simple_waypoint.py

+                    command = self.goals.pop(0)
+                else:
+                    command = commands.Command.create_land_command()
+            case drone_status.DroneStatus.MOVING:


You don't have to do anything while the Drone is moving, we just let it finish it's last command. The loop is for making fine adjustments when on top of the landing pad.

BalajiLeninrajan · 2024-09-25T21:12:45Z

modules/bootcamp/decision_waypoint_landing_pads.py

+                    )
+                else:
+                    command = commands.Command.create_land_command()
+            case drone_status.DroneStatus.MOVING:


You don't have to do anything while the Drone is moving, we just let it finish it's last command. The loop is for making fine adjustments when on top of the landing pad.

Just made another commit, I'm unsure what you meant by making fine adjustments (perhaps I am supposed to reroute to the destination if position != destination)? I've left it for now just to resolve the other issues first.

BalajiLeninrajan · 2024-09-25T21:13:56Z

modules/bootcamp/decision_simple_waypoint.py

+                    command = commands.Command.create_land_command()
+            case drone_status.DroneStatus.MOVING:
+                # if the current position is close enough to the destination.
+                if (


Check this when the Drone is halted.

BalajiLeninrajan · 2024-09-25T21:17:15Z

modules/bootcamp/detect_landing_pad.py

        bounding_boxes = []
+        for rectangle in boxes_cpu:
+            optional_bounding_box = bounding_box.BoundingBox.create(rectangle)


Deconstruct the tuple result, box = bounding_box.BoundingBox.create(rectangle)

BalajiLeninrajan · 2024-09-25T21:17:43Z

modules/bootcamp/detect_landing_pad.py

+            optional_bounding_box = bounding_box.BoundingBox.create(rectangle)
+            # why is it that this one returns False while other functions might return None for a failure?
+            # still unsure why .shape() is required
+            if optional_bounding_box[0]:


Check for failure and return an empty array if that is the case

BalajiLeninrajan

You're almost there! You just need to sort out some stuff with the loop logic. I you don't understand what I mean you can shoot me a msg on Discord or find me at the weekly auto meeting.

BalajiLeninrajan · 2024-09-26T13:34:26Z

modules/bootcamp/decision_simple_waypoint.py

@@ -70,6 +87,17 @@ def run(

        # Do something based on the report and the state of this class...

+        if report.status == drone_status.DroneStatus.HALTED:
+            if self.goals:


I don't really see why you have a list of "goals". You can simply check if the drone is within the acceptance radius, land if true, set relative destination if false.

BalajiLeninrajan · 2024-09-26T13:34:39Z

modules/bootcamp/decision_waypoint_landing_pads.py

@@ -70,6 +119,31 @@ def run(

        # Do something based on the report and the state of this class...

+        # only execute when "drone" is ready for another instruction
+        if report.status == drone_status.DroneStatus.HALTED:
+            if self.goals:


I don't really see why you have a list of "goals". You can simply check if the drone is within the acceptance radius, land if true, set relative destination if false.

BalajiLeninrajan · 2024-09-26T13:36:32Z

modules/bootcamp/detect_landing_pad.py

+            else:
+                # return an empty list if any of them fail???
+                bounding_boxes.clear()  # clear the boxes if any were made before the failed one
+                break


This solution works, but consider that

if not result: return [], image_annotated

achieves the same result with less code (and slightly more performance since you don't have to clear bounding_boxes). (You do not need to change anything here, just something to think about).

Mhm, this one was a bit of a 50/50 choice for me. I was thinking "if the flow goes to the return statement anyways, I'll just make sure the return tuple is correct then return". HOWEVER, I like the idea of better efficiency :P

Hardcoded the decision tree for task #3 and task #4, this allows for simplicity and readability Switched the clearing of the list to an early return statement

BalajiLeninrajan

Super close, you should be able to finish this time around!

BalajiLeninrajan · 2024-09-27T15:50:47Z

modules/bootcamp/decision_simple_waypoint.py

+                command = commands.Command.create_set_relative_destination_command(
+                    self.waypoint.location_x, self.waypoint.location_y
+                )
+                self.landing = True


You're super close, you have to set this to true only when the drone is within the acceptance radius, you code will look something like

if drone is halted: if drone is within acceptance radius: land the drone else move the drone

BalajiLeninrajan · 2024-09-27T15:55:24Z

modules/bootcamp/decision_waypoint_landing_pads.py

@@ -70,6 +113,34 @@ def run(

        # Do something based on the report and the state of this class...

+        # only execute when "drone" is ready for another instruction
+        if report.status == drone_status.DroneStatus.HALTED:
+            if not self.finding_landing_pad:


Similar to task 3, super close just some minor changes in the conditionals:

if drone is halted: if drone is looking for waypoint: if drone is within acceptance radius: get nearest landing pad and tell the drone that it should look for it next else: move the drone else if drone is looking for landing pad: if drone is within acceptance radius: land the drone else: move the drone

BalajiLeninrajan

You just need to improve your variable names a bit and you should be done with the bootcamp!

BalajiLeninrajan · 2024-09-27T21:26:35Z

modules/bootcamp/decision_waypoint_landing_pads.py

+
+    @staticmethod
+    def calculate_distance_squared(
+        location_1: location.Location, to_x: float, to_y: float


Use more descriptive variable names

BalajiLeninrajan · 2024-09-27T21:26:51Z

modules/bootcamp/decision_waypoint_landing_pads.py

+
+    @staticmethod
+    def get_relative_position(
+        location_1: location.Location, location_2: location.Location


Use more descriptive variable names

BalajiLeninrajan · 2024-09-27T21:27:07Z

modules/bootcamp/decision_simple_waypoint.py

        # ============
        # ↑ BOOTCAMPERS MODIFY ABOVE THIS COMMENT ↑
        # ============

+    @staticmethod
+    def calculate_distance_squared(location_1: location.Location, to_x: float, to_y) -> float:


Use more descriptive variable names

BalajiLeninrajan

Lgtm.

bantuerfei0 added 5 commits September 24, 2024 12:17

Task #1

8d7d5d0

Possible missing error handling and redundancy

Task #2

fc218e2

Task #3

878401f

Task #3 tweaks and Task #4

b201dd3

Linting

db67208

BalajiLeninrajan reviewed Sep 24, 2024

View reviewed changes

bantuerfei0 added 3 commits September 24, 2024 19:34

Generalized pathing for Task #3 and Task #4

3215413

Improved Task 1, 3, and 4 based on feedback

Linting

ab66aba

Add docstring

44bad08

bantuerfei0 requested a review from BalajiLeninrajan September 24, 2024 23:48

BalajiLeninrajan reviewed Sep 25, 2024

View reviewed changes

Attempted to fix misunderstood instructions

519532c

BalajiLeninrajan reviewed Sep 26, 2024

View reviewed changes

bantuerfei0 added 2 commits September 27, 2024 09:01

Apply suggestions

9f7df4d

Hardcoded the decision tree for task #3 and task #4, this allows for simplicity and readability Switched the clearing of the list to an early return statement

Linting

ac09940

bantuerfei0 requested a review from BalajiLeninrajan September 27, 2024 13:08

BalajiLeninrajan reviewed Sep 27, 2024

View reviewed changes

Implemented structure changes

1c24b3d

bantuerfei0 requested a review from BalajiLeninrajan September 27, 2024 16:21

BalajiLeninrajan reviewed Sep 27, 2024

View reviewed changes

bantuerfei0 added 2 commits September 27, 2024 19:05

Renamed function parameters to be more descriptive

302a4cd

Linting

26ebc01

bantuerfei0 requested a review from BalajiLeninrajan September 27, 2024 23:06

Added type hinting

c7f887a

BalajiLeninrajan approved these changes Sep 28, 2024

View reviewed changes

BalajiLeninrajan closed this Sep 28, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Alex Liu - Autonomy Bootcamp 2023 #105

Alex Liu - Autonomy Bootcamp 2023 #105

bantuerfei0 commented Sep 24, 2024

BalajiLeninrajan left a comment

BalajiLeninrajan Sep 24, 2024

BalajiLeninrajan Sep 24, 2024 •

edited

Loading

bantuerfei0 Sep 24, 2024

BalajiLeninrajan Sep 24, 2024

BalajiLeninrajan Sep 24, 2024

BalajiLeninrajan Sep 24, 2024

BalajiLeninrajan Sep 24, 2024

BalajiLeninrajan Sep 24, 2024

BalajiLeninrajan left a comment

BalajiLeninrajan Sep 25, 2024

BalajiLeninrajan Sep 25, 2024

bantuerfei0 Sep 26, 2024

BalajiLeninrajan Sep 25, 2024

BalajiLeninrajan Sep 25, 2024

BalajiLeninrajan Sep 25, 2024

BalajiLeninrajan left a comment

BalajiLeninrajan Sep 26, 2024

BalajiLeninrajan Sep 26, 2024

BalajiLeninrajan Sep 26, 2024

bantuerfei0 Sep 26, 2024

BalajiLeninrajan left a comment

BalajiLeninrajan Sep 27, 2024

BalajiLeninrajan Sep 27, 2024

BalajiLeninrajan left a comment

BalajiLeninrajan Sep 27, 2024

BalajiLeninrajan Sep 27, 2024

BalajiLeninrajan Sep 27, 2024

BalajiLeninrajan left a comment

Alex Liu - Autonomy Bootcamp 2023 #105

Alex Liu - Autonomy Bootcamp 2023 #105

Conversation

bantuerfei0 commented Sep 24, 2024

BalajiLeninrajan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

BalajiLeninrajan Sep 24, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

BalajiLeninrajan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

BalajiLeninrajan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

BalajiLeninrajan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

BalajiLeninrajan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

BalajiLeninrajan left a comment

Choose a reason for hiding this comment

BalajiLeninrajan Sep 24, 2024 •

edited

Loading