As I am getting closer to our first RoboMaster System One tournament in Montreal, and after interacting with quite a number of people with Chip and Dale, I am wondering if DJI missed out a major aspect of the product rollout of the RoboMaster S1 and whether that will ultimately snuff the robot off the retailers' shelves.

So, I am asking you all on this forum: What are the hurdles you are facing when trying to get people to adopt and go out to buy a RMS1 (either for themselves, or for their kids)?

Since most people think it's a toy, the price is the first hurdle.  Someone has to be into robotics and/or artificial intelligence to really appreciate the S1 and to justify the price.  And I don't thnk DJI marketing reach the mainstream media/audience.  The market is there and a lot of hobbyist will buy the S1 if they only knew about it and it's potentials.  I'm really surprised that there's not a lot of S1 owners in Southern California.

Hello and good day rhoude57 - YUL. Thank you for reaching out and for these information you have shared with us today. I hope that you will get the best information from our valued DJI members with regards to this inquiry. In addition I will cascade this thread to the proper DJI department for further attention with regards to this issue. Thank you for your support.

  

I have a lot of colleagues that are really interested in the S1, but won't buy it at the moment. Let's face it, at its current stage the S1 is just a nice & fun RC Toy with limited programming abilities. The sandbox architecture hurts more than it helps and keeps the advanced target audience away. You can get more flexible stuff at a cheaper price tag and a lot of people I know don't want to spend (or have) time to reverse engineer the S1 to get more out of it.

This is from the DJI product page:

In my opinion this is false advertisement or raising at least totally wrong expectations.

In my spare time I'm teaching Scratch & Python for talented/interested grade schooler & freshman. From an educational point of view the S1 is (among other things) way too expensive. I can't setup classrooms for the S1 (there is no teacher/student functionality at all), every student would need an additional DJI account (we are using PCs with very limited Internet access, GDPR is a nightmare where I come from) and the standard S1 software is just impractical for teaching.

All the bleeps, sounds and pings from the software will drive you crazy in a classroom situation. The standard .DSP project files are AES encrypted, we found a solution for this (thanks BGA), but that's not the point. I'm not teaching hacking or reverse engineering classes...

In 1:1 with my eight-year-old the S1 is working pretty well

As always: This is just my personal opinion. You may not like it, but that's how I see the S1 atm.

Cheers

  

I think I would mostly agree. Even though the engineering is superb, the price does seem high.

You can do some interesting things with the programming environment, but you quickly run into roadblocks. I wish the UART connection was documented.

The blaster seems to be the main reason why some functionality is locked out. Without restrictions, the device would easily be able to find people and shoot at them, and this would lead to some very bad publicity! Hence the almost paranoid attempts to prevent hacking.

You can find Chinese suppliers selling a complete robot kit with RPi + LIDAR running ROS for about the same price.

sevreNniarB Posted at 11-16 08:59
I have a lot of colleagues that are really interested in the S1, but won't buy it at the moment. Let's face it, at its current stage the S1 is just a nice & fun RC Toy with limited programming abilities. The sandbox architecture hurts more than it helps and keeps the advanced target audience away. You can get more flexible stuff at a cheaper price tag and a lot of people I know don't want to spend (or have) time to reverse engineer the S1 to get more out of it.
This is from the DJI product page:
[view_image]

"The sandbox architecture hurts more than it helps and keeps the advanced target audience away."
Agreed!
The S1 has a UART and SBUS port with no way of interacting with these ports. At one time DJI's FAQ page said using these ports voided the warranty.
I personally would think the price for the quality of parts is reasonable IF one had full control of all of the hardware.
If the S1 were more open, I'd be quick to recommend it to others.
The photo showing a robot arm on a S1 chassis does appear to be false advertising. It likely wasn't intentionally false. The person who put together the robot probably thought the SBUS port could be used as an input.
Just access to the SBUS and UART would be a huge step in making the robot more versatile. I'd also like the option of powering the S1 from other batteries without having to reverse engineer the Intelligent Battery's protocol.

Agreed！Gimbal and blaster limits the development and expansion of s1. If all control interfaces are open and functions can be combined at will, then s1 is still a good educational product.

Duane Degn Posted at 11-16 12:25
"The sandbox architecture hurts more than it helps and keeps the advanced target audience away."
Agreed!
The S1 has a UART and SBUS port with no way of interacting with these ports. At one time DJI's FAQ page said using these ports voided the warranty.

Agreed

sevreNniarB Posted at 11-16 08:59
I have a lot of colleagues that are really interested in the S1, but won't buy it at the moment. Let's face it, at its current stage the S1 is just a nice & fun RC Toy with limited programming abilities. The sandbox architecture hurts more than it helps and keeps the advanced target audience away. You can get more flexible stuff at a cheaper price tag and a lot of people I know don't want to spend (or have) time to reverse engineer the S1 to get more out of it.
This is from the DJI product page:
[view_image]

Agreed

In my recent experience with the organization of what would have been the first RoboMaster System One - Montreal event, I found that many are intrigued by the RMS1, but not many are willing to dish out $500US dollars without giving the robot a spin. On the other hand, existing owners are a little queazy letting strangers messing around with the robots and their personnal smart phone or PC on top of that.
Also, many are intrigued by the competition aspect of the robot, but are left on their appetite by all the review videos available on the Internet.
What is needed, IMHO, is a tournament kit provided by someone (first thought being that this should be DJI) and some support from authorized resellers with making a venue and the tournament kit available for leagues to emerge. The reseller could make some "loaner" RMS1s available for try-outs. This would also be a great marketing opportunity for those resellers.

Yep I agree with all the above - my main obstacle being this robot is so blind - we have no way to give it any sensory input like lidar to do any basic obstacle avoidance.

Plus we have no access to the video stream - for which there are some cool Python modules that you can figure out to process the video data and make decisions in your programs.

I don't miss having a claw or scoop to move objects around - that's fine for other robots.

The S1 is a gun-toting light battle tank robot and needs sensory input and comms.  

Think "Ghost in the Shell".  The S1 would be a fine "Think Tank" for "Section 9" in-city operations.
They were called Tachikomas, and they had a collective A.I. that could allow them to learn from each other:

MarkusXL Posted at 11-19 07:00
Yep I agree with all the above - my main obstacle being this robot is so blind - we have no way to give it any sensory input like lidar to do any basic obstacle avoidance.

Plus we have no access to the video stream - for which there are some cool Python modules that you can figure out to process the video data and make decisions in your programs.

They did a pretty good job with the Tello EDU: You can just send UDP commands from your PC, use every library in the world, ...

The S1 users have to reverse engineer network traffic, use arbitrary code execution and we are not even close to the things that are possible with a Tello...

Cheers

sevreNniarB Posted at 11-19 11:32
They did a pretty good job with the Tello EDU: https://www.youtube.com/watch?v=RHRQoaqQIgo You can just send UDP commands from your PC, use every library in the world, ...

The S1 users have to reverse engineer network traffic, use arbitrary code execution and we are not even close to the things that are possible with a Tello...

I totally agree! The DJI Tello and Tello EDU were deployed with a complete, well documented SDK.
This has allowed other entrepreneurs to develop alternative control apps for the drones, such as DroneBlocks.
I would have expected that the RoboMaster S1 would have been deployed not ony with a similar SDK, but with the ability of meshing S1s and Tellos, pretty much like the real RoboMaster eco-system.

Yes I'd agree the price is rather high, and that is why some interested people are holding off, but once you realize all the technology packed in the RMS1, I'm tempted to say the price is justified. In hindsight, DJI could balance a bit more cost vs. adoption rate.

Now, I've been scanning all the BlackFriday ads available to date, and only one of them has a Spark drone for $250. I was hoping to see more, especially the Mavic Mini and RMS1 ... but somehow these are nowhere to be found. It's as if DJI wants their products to remain a 'niche' product i.e., not for the masses. DJI should hire some Marketing Executives from Apple ...

MarkusXL Posted at 11-19 07:00
Yep I agree with all the above - my main obstacle being this robot is so blind - we have no way to give it any sensory input like lidar to do any basic obstacle avoidance.

Plus we have no access to the video stream - for which there are some cool Python modules that you can figure out to process the video data and make decisions in your programs.

...my main obstacle being this robot is so blind - we have no way to give it any sensory input like lidar to do any basic obstacle avoidance.

That's why I will never buy this robot
others see it as similar

Greeting Andre

VJS-Tiger Posted at 11-22 07:28
...my main obstacle being this robot is so blind - we have no way to give it any sensory input like lidar to do any basic obstacle avoidance.

That's why I will never buy this robot

And I can understand the frustration, especially when DJI has everything in place, or just about, to make the RoboMaster S1 a great robotics platform.

Imagine the possibilities if DJI released a Manifest 2 equivalent to the Jetson Nano (i.e. a Manifest Nano, you say???), with the ability to interface and exchange with the RMS1 Intelligent Controller.
Then, add to that a stereo vision system augmented with a built-in IMU (forget LIDARs... they are so "passé")...

JB63 Posted at 2019-11-21 08:51
Yes I'd agree the price is rather high, and that is why some interested people are holding off, but once you realize all the technology packed in the RMS1, I'm tempted to say the price is justified. In hindsight, DJI could balance a bit more cost vs. adoption rate.

Now, I've been scanning all the BlackFriday ads available to date, and only one of them has a Spark drone for $250. I was hoping to see more, especially the Mavic Mini and RMS1 ... but somehow these are nowhere to be found. It's as if DJI wants their products to remain a 'niche' product i.e., not for the masses. DJI should hire some Marketing Executives from Apple ...

I looked into it.   If you priced the individual components, such as the mecanum wheels, motors with built-in ESC, 2 axis gimbal, you'll spend close or even over the price of the RMS1.  The wheels alone would probably retail $40-$50 each.

But it would be nice to have more control over the system both at higher and lower levels.
By lower levels, I mean as an example being able to control the wheel motors individually so one could experiment with the the physics behind the technology.

I'm very new to the S1, and, as I've recorded here on this forum, have had some difficulties with it.   I believe the firewall for new owners features:
A: Cost - it's damned expensive
B: Software - a quick glance at the interwebs will easily advise potential users that therein lies the weak spot.  Dodgy, buggy and weak. Python largely undocumented, Scratch a steep hurdle. Only today I learned that under Scratch you can only blast beads, not fire the laser, for example.  Python programmable? Sure, but where's the documentation? It's an odd implementation of a language I'm very adept with.
C: This forum - look at the DJI drone forums, hundreds of posts.  Look at this forum, pretty sparse.  How many of us S1 owners are there? A casual visitor might think there's about 17.

I think we have a highly capable device but I find myself unable to say 'get out there and get one!'

CallMeAlan Posted at 1-6 10:10
I'm very new to the S1, and, as I've recorded here on this forum, have had some difficulties with it.   I believe the firewall for new owners features:
A: Cost - it's damned expensive
B: Software - a quick glance at the interwebs will easily advise potential users that therein lies the weak spot.  Dodgy, buggy and weak. Python largely undocumented, Scratch a steep hurdle. Only today I learned that under Scratch you can only blast beads, not fire the laser, for example.  Python programmable? Sure, but where's the documentation? It's an odd implementation of a language I'm very adept with.

The development documentation is here:

https://www.dji.com/robomaster-s1/programming-guide

This reminds me that it seems there have been some silent updates to the guide. At least I am seeing new stuff there.

BGA Posted at 1-6 10:17
The development documentation is here:

https://www.dji.com/robomaster-s1/programming-guide

In many cases, the updates probably covered existing but poorly or undocumented features.

Six months now since the RMS1 hit the streets, I still believe DJI fumbled the RMS1 product launch. The software was clearly not ready for a roll-out and neither were the production and delivery lines. That placed the software development team on their heels and they still have not recovered.

There is a lot of great hardware under the RMS1 hood. I think the product would turn out to be a much greater success if DJI had rolled out an open and well-documented SDK/API and let hobbyists bring out the capabilities out of the machine.

I always use the DJI Tello and DroneBlocks as a perfect example of the opportunities a good SDK/API open up.

Not to start the new year on a negative note ... but I'm starting to get disappointed with the RMS1 ... After assembling it and playing with it in RC mode, it's already getting 'passé'. COuld be my attention span ... but come on DJI, the machine is very capable, yet all we see are non-sensical limitations.

No Linux support is a bummer :/ I'd love to have an intergrated webserver in the bot which mimics the app so basaically all you need is a laptop in the same LAN to control the bot.
A fully documented REST API would be another way to solve that, people would create their own control software for every OS in no time.

I agree with all the comments on this thread.  It would be lovely if they just open sourced the platform.  Then we could pick up their slack in terms of tooling and potential power.

fk6 Posted at 1-8 11:53
No Linux support is a bummer :/ I'd love to have an intergrated webserver in the bot which mimics the app so basaically all you need is a laptop in the same LAN to control the bot.
A fully documented REST API would be another way to solve that, people would create their own control software for every OS in no time.

An unannounced text mode API is in the works (in fact, I can use it in my robot ;) ). Other than that, my code (although it is far from being very useful) runs fine on Linux: